DD

I was a PE on the TPE team within the larger Traffic organization at FB, and I thought about the security, reliability, and scalability of FB's edge platform every day for 5 years. I spent a lot of time tackling super low-level problems, and I took these challenges as opportunities to deliver big wins, grow my team, publicize my team's accomplishments, etc. I became quite adept at translating context between various classes of stakeholders, including ISPs, software engineers, and network contractors.

Prior to that, I received a masters-level CS education from Rutgers University in NJ, with a major focus on Algorithms, AI, and Distributed Systems, AI, and a minor focus on Probability and Astrophysics. I spent 3 years leading student organizations and running HackRU, the first-ever and 2nd-largest student hackathon. And through that involvement, I also became quite involved with a nonprofit in NYC called hackNY which runs a fellowship program that brings together 20-30 brilliant, responsible-minded technologists to live and work together in the heart of NYC, and inducts them into a larger community of hundreds of alumNY. I have been directing this program for the last 2 years.

- 5 years ago, I was 22, still a student. Now, I'm 27, I'm a homeowner, DM, nonprofit director, senior-level SRE. 5 years from now, I'll be 32, and I'd love to be leading an eng org. How do I get there? What has your own career journey looked like?

MONGODB IV: Manager: Ernie Hershey
- What have been your career highlights at MongoDB? ... What are you looking forward to the most, and how are you still growing? How much longer are you expecting to stay?
	- I've been thinking about that. I don't know how long I'll be here. I'm not an expert on the ins-and-outs of other people's growth. He's hired 10 people.
	- We've had minor layoffs in the last year. Akshay is his boss, but isn't checking up on him. He's learning new tech, "I guess". He runs a reading group.
	- There's been bad news on the earnings call in the 10.5 years he's been there. He thinks of it as a stable trajectory, not a rocket ship.
- I'd like to get behind the wheel of this rocket ship. What are your expectations from a Mid-Level SRE, from a Senior SRE, from a Staff SRE, etc.?
	- They have 3 shapes of SRE well-defined. Staff SREs get to Staff because they care about higher-level problems like what matters to the company, what moves the needle.
	- Staff: Forging relationships with teams that are far away from SRE. Getting things done with the Kube platform, talk to the business people, Cloud people, etc. And make SRE part of the solution to these big problems.
	- Senior: A smaller version of that. Good relationships within SRE, seeing and prioritizing problems within your lane, tea-leaf reading, etc.
	- Mid-Level: Foot Soldier, doing the things that are directed to you.
	- We're good at promoting people and a lot of that is on leadership to identify people's strengths. This is less tea-leaf reading and more process-driven: Development Plans.
	- They have an above-average review system. They used OKRs for a few years (stolen from Google).

- Cailin talked to me about some of the XFN projects they're excited about: turning Atlas Serverless into a truly serverless product, decoupling the Server Binary, the smooth allocation of nodes to client VPCs, etc. (Does the Special Projects) Team have stake in this? What is the roadmap and how does it fit into the (Cloud) Org's Roadmap?
	- I admire Cailin a lot - she's an Architect at heart. The SREs are the biggest stakeholders in the Architecture of Cloud. Clarification: Is there overlap between Team/Org Roadmaps?
	- He wants an SRE-wide conversation about improving the Oncall Experience. He has 2 peers: Directors, who are less interested in that conversation because their oncall is segmented by Product. "We" can get paged for problems with Atlas, any other products, and also for FedRamp and Vuln Mgmt. We can't ignore any segment of the entire cloud environment because we are responsible for all of its breadth in special environments. So folks will come to our conversation and learn about how to improve their oncall rotation.
	- There's a big project right now that scans for vulnerabilities in container management. Everyone in Cloud will have a stake in this.
	- It's not quite as XFN as the stuff Cailin's thinking about, but we are XFN with a lot of teams in Cloud.
- I'd like to get a better sense of the (Special Projects) Team Roadmap again. Where do you think I can fit in? How long is onboarding? When do I go oncall? How often are meetings?
	- They don't have an "artifact" that renders their Roadmap. They plan for quarters. The whole department only plans for the next quarter.
	- Scopes are Project-Level, they're pulled into JIRA, and those are filtered into Quarterly Plans. Examples: ECR Migrations out of AWS GovCloud, Kube Migrations, Argo, etc.
	- Onboarding is 90 days (3 months). Oncall could be before or after that. Oncall is actually very easy - too easy. No one's complaining, but he's uneasy - 2-3 non-crit pages/day.
	- He feels uneasy because a lot of their work is predictive of failure. They're moving FedRamp into Azure, they're going to scan 3rd-party images, etc. This quiet period won't last.
	- He's excited for me to join because I'm good at coding but less so at non-coding. He's basically an old Sys-Admin and likes that I have a good eye for Building. Sees a synergy.
	- He wants to see a Builder's solution to the problems in the infrastructure.

MONGODB III: EVP of Cloud: Cailin Nelson
- Talk to me about the work you're doing. What does it look like on a day-to-day?
	- My life is not as fun as it used to be. She's working on an Acquisition, org changes, building an ML Ops team.
	- 2 VPs are on sabbatical so she's been checking on their directors.
	- What I LIKE doing is reviewing design docs for XFN projects.
- Talk to me about your personal convictions. What are the technical/nontechnical endeavors that get you most excited? What are MongoDB's big bets right now?
	- Single biggest question now is building a Serverless Cloud "MongoDB". Right now they have "Atlas Serverless" but it's mostly smoke-and-mirrors.
	- It's not an easy problem. MongoDB is open-source, and always came with horizontal scaling, and always designed to be simple.
	- We need to decouple the Networking / Query Planning / Query Execution / Storage layers from the binary.
	- 2 mistakes to dig out of. In order to give Atlas private networking, we put each customer in one VPC and use functionalities of each Cloud Providers.
	- Customers like that they are in their own VPC. But this makes it tough - it takes time to startup a new node. It would be cool to have standby nodes easy to allocate across customers, but to move nodes from VPC to VPC, they have to be shutdown and restarted. How do we dig ourselves out of this?
	- We have OUR app service tier running across all Cloud Providers. We have to put OUR IPs in each customer VPCs. Cloud Providers limit the amount of IPs to 400 - we have 100 - so we have a Proxy Tier that provides our app services over 2 IPs. There's an open source product called Lighthouse to help coordinate this.
- Talk to me about the work on the Special Projects team with Ernie. Where will it be in 6mo, 1yr, 2yrs?
	- Cloud Team vs Core Server Team (DB). We'd like to tear down the boundaries though. Only she and one of her VPs care about data locality. She's running pilots - some move people and some move responsibilities across the boundary. Instead of MongoDB server binary.
- I'd like to get a better idea of your own career journey - why have you stayed at MongoDB for 10 years?
	- The team: There's been a lot of people who have been there for 10 years. The thought of not working with them is distressing.
	- It's impossible to get bored. It's a large-and-complex enough product. I've done a lot of consumer-facing products holistically and found it's more fun to NOT do so.
	- There's a level of user-sophistication to count on. The Cloud team was 6 people when I joined, soon managing 11 DRs - 450 people with 30 SREs now.
	- We used to just whiteboard in a conference room and do stuff. We had to formalize that as we grew. 1-2 pagers called a Scope.
	- Operationally, it was hilarious. Akshay and others carried some of the 12 servers in the office with disk failures and such. No automation.
	- 450k VMs now in the Cloud managed by 450 folks. We're really good at automating failures away. Soft boundary between SWE/SRE - customers can only connect to DBs.
	- It's not managed by Chef or Puppet. There is a chef-solo layer, but most of it is self-diagnosis Java code.
	- Some boundaries have hardened. A salesperson used to be able to crash her day, but no longer. But it was nice when I knew everybody's name.
	- I was famous for knowing what everyone worked on at the 80 people level.
- As _, what are some of the major projects and big bets you've worked on / led to fruition?
- How are you being intentional about your culture - what are you looking for in each new engineering hire?
- Something you appreciate and something you wish was different?

POLYMER IV: Technical
- Embedded SWE, into AWS Virtualization, into Twilio 3 years, doing Google Meet.
- This guy got pulled into it by Bo. "This guy's trying to scam me." Then he talked to Derek.
- "I didn't mind the idea of joining a startup." He was done with the political BS at Twilio.
- The team is the most senior he's been with by far. Everyone has a lot of experience.
- Engineers tend to be more independent. What does your day look like? How does it bring you to market?
- He's putting a public testnet in the cloud. He's putting light clients and relayers in the cloud.
- ETH forked and set up new data structures. So he spent time thinking about backwards compatibility.
- Daily life: contributing to ETH with 5k line commits. Next week it should be landed. He starts and stays very late.
- They do an all hands every Monday. Typically people keep to their work streams, but they communicate over Discord often.
- Weird naming conventions. "Smart Contract : EVM Bytecode" "Root Computation : Hash"
- The "Twilio" political problem is Hierarchical. The VP was an asshole and a total idiot, and hated his manager, who was terrible at communication.
- Also, he was established as a "Champion" i.e. TL but no benefits or voice. Dealing with the open source community is not personal, it's just technical.
- Do you feel like this silos you into web3? Most web3 companies are scams. He doesn't feel like he's doing "web3 work".
- It feels like magic, but it's just code. It's really cool technology, though. The difference is a change in mindset - "Who will run this relayer?" "Whoever."
- He doesn't mind making changes in his life. It all boils down to transistors and code in the end anyway. Build many different skills not tied to blockchain.
- i.e. Go, TypeScript, Consensus Algorithms, ZK cryptography, Rust... It won't close any doors, but you won't get the MongoDB flag. "Best place to be is a good startup right now."
- "I probably would have chosen AWS over Polymer. It was early in my career and I needed a CV boost." "You don't need another CV boost after FB."
- S Corp? He's in Spain, so the laws are different. If you work locally, you're passive. If you're "autonomous", you "sell" entities your time.
- It sounds like a lot of work, but it isn't. He hired a paperwork person, but wishes he didn't. He just collects his paystubs and submits them.
- Most are suspicious when they hear you have to self-employ, but he said "oh I can save taxes!" He can pay off his car, phone, laptop, etc. You can't write off everything though.
- They get a 25k health insurance card from Ramp. Also free lunch, $100/week - new info. Good doctors are in public health insurance.

POLYMER III: Nontechnical
- Joined Polymer 1.5mo ago. Working for 10 years. Don't join a company that you're not sure about.
- 2 vectors: role itself, corporate development, risk assessment, operationalizing acquisitions from Twilio/Coinbase.
- 1st vertical: role Their choices were web3 director/head-of-ops, staying in the space
- 2nd vertical: company is well-funded with strong investors. looking for very early company: high growth, high conviction, etc. believe in the founders - slack used to be gaming. believe in the team - heavy hitters, go getters.
- don't put a lot of stock into the token equity, because it's not the same as stocks. most web3 companies have "utility" and "governance" tokens. social capital drives token value. they're not a simple SaaS company.
- comes from Coinbase corporate development so the craziness of managing crypto is as hectic as you think. she spoke to people with conviction: Bo, investors from Bcap, etc. She heard the areas of concern and said "that's where I can help" - specifically "bringing us to market" - they've spent a lot of time doing R&D but less on marketing. she's convinced that they have the best tech, but is unsure whether they can get everyone to Care. ex: we're android, but need to convince people we're apple. IBC gives us credibility, but we need to be thoughtful of the story we're telling. It's huge value, and someone has to do it.
- shifted to Coinbase because they needed M&A - it's like web2.5. What value is it delivering to people today? I have no answer. In the last bull-runs, there was Crypto, then NFTs, and now it's slowing down - the industry is figuring out use cases now: authorization, authentication, etc.
- crypto for institutions: first use-case. crypto is slowly maturing into a widely adopted currency in institutions. "building the ship while we're flying it".
- stay in web2 to build a name brand. 5 years is "just enough" but it's also "minimum just-enough". if you join a startup and make a name for yourself in it, that is very credible. it's very hard to invest in first-time founders, but Peter and Bo are very investible.
- More engineering mentorship - go to MongoDB. More business/product into ecosystem - go to Polymer. Stripe didn't get a product person until employee 200 - make our engineers product-minded.

POLYMER II
mentorship
meet people from the board/investors
safety, return, and mentorship
how does s-corp work? mental overhead?
10-20x? is that right?
it's the time in my life to build a moat of seniority, and also a time to take risks.
50k more salary, no question
technical mentorship vs industry mentorship

comp? 10-20x increase? "why would I ever write an email? AI? EVs?" the adoption cycle of innovation is long - early, late, laggards, etc. We're still in early adopters.
27k crypto devs. 27m devs in the world. 0.1% of the dev population since 2009. ETH didn't kick off until 2016, which started real development.

org leadership
mentorship?
board/investors - token2049 is next week, so its tough to set up meetings. but I can talk to one eng and one non-eng and watch an earnings call.
text Eddie my telegram
he doesn't have an S-Corp. He still has to pay taxes for last year.
Oct 2-6 in LA.

The richest people in the world save the most money on their taxes. We're learning about this in our late 20s set up our own S-Corps. They are continuously re-structuring.
If you want to be 1/10th of an engineering team, join us.

What sounds the most interesting to me? Career growth. The sky is the limit with comp. Sees me collaborating on ZK research (cutting edge cryptography) or protocol work.
An EM wants to move to coding, so I could fit into that role.
Also, we need documentation - there's a single person leading DevRel documentation.
As founders, they can give me more cards to move beyond my work at Polymer and into the space.
Can create my own company down the line. I could be leading the people on the Team page - these are titans in the industry. They brought non-crypto people into crypto.
Transferability: They're building core infra - the connections between blockchains, not blockchains themselves.

POLYMER:
- Future is pretty open-ended. Won't be trapped in DevOps, path to leadership in an angle of my own choosing.
- Compensation - 150k cash and 170k token. Company building in web3 - social capital concentrates at the top - staff and principal.
- In the DeFi space, the devs are the stars - the social capital is most famous in the company. i.e. Yearn devs got the most social capital.
- Priority for the org is to build the brand of the individual devs. Will send me to a lot of conferences to help me build my personal brand.
- Translations to other spaces? Pivots are not that much of an issue. Infra problems are infra problems regardless of vertical - are you the one giving talks, publishing work, etc?
- Reasonable Expectation is minimum 4-5x, 10-20x, increase in token equity after token minting.
- Everyone does DevOps, so I will not be alone. Everyone in engineering is heads-down coding. No one is armchair QBing.

- Peter from Polymer asked me what I see myself doing in 5 years. I thought about it, and said I'd like to lead an engineering org by 32.
- My brother and I spoke about this and we established a consensus that in order to do so, I need to be building my WIS score now - by not just managing technical projects, but managing teams, having access to incredible SRE mentors, delivering talks at public conferences, and building my reputation across the industry.
- I'm thinking about this decision from a 2-year timeframe.

Tough Decision Tree
- Datadog 255k - Frankly the offer is relatively low. It's all liquid, but I'm not particularly excited by the work, and I'm not convinced I will find opportunities for career growth.
- TikTok 330k - Great TC package. But I'm concerned about the liquidity, culture, and longevity (maybe it will be banned before it goes public).
- Polymer 280k - I love the work, the team, and the problem it's tackling - I'm very excited about the future of Polymer. I'm a little concerned about the liquidity/TC, and mostly concerned about the fact that it will and must remain a small team (though I understand why), and about the lack of staff+ SRE mentorship that would be available to me. In the worst case, I would still get solid TC and learn a lot about running a business in the web3 space. Problem is, I don't know if I want to stay in the web3 space in 5 years.
- MongoDB ???k ~300k? - It's all liquid. I'd be Senior at a public tech company with a great reputation across the wide tech industry, which I think is important at this stage in my career, if my desire is to lead an engineering org by 32 - building a "moat" of seniority around me, which could be considered as important as building a Savings account. Worst case, I get solid TC (fully liquid) and have the opportunities to grow as an engineering leader in the tech industry, by ways of public talks, access to staff and principal SRE mentorship, etc.

Decision Tree:
Lowest Rank: Datadog
OK, good TC: TikTok
Polymer and MongoDB come in - probably MongoDB. I have to think about it.
Polymer comes in - Polymer
MongoDB comes in - MongoDB
Neither come in - TikTok

14-15 INT, 12-13 WIS, 14 CHA
pump WIS to be director by 32. need public talks and direct mgmt

outline life aspirations
update personal website
add interviewers to linkedin

MONGODB

mongodb vp of sre
- he's probably very good - 11 years to VP. he's been growing with the company. emotionally intelligent + technically competent.
- avoid: scatterbrained thinking. be thoughtful and comprehensive. come with narratives prepared. he wants to see high potential.
- he's probably also a loyal control freak. there's probably 100 SREs and 50+ seniors. be calm and controlled. he probably assigns HC.

Questions
- As VP of SRE, what are some of the big bets you've worked on / led to fruition?
- I'm 27 - 5 years ago, I was still a student. Now, I'm a homeowner, DM, nonprofit director, senior-level SRE. 5 years from now, I'd love to be leading an engineering organization. What has your career journey looked like?
- What are the technical or nontechnical endeavors that get you most excited? What are you personally looking forward to at MongoDB?
--> Moving to a new platform (EKS). Aim is to get this completely self-serve - stateless workloads. Simplifying the Kube world. Security work on containers. Perf. Mostly interested in establishing a Socket in the Sky, charge in a predictable way for storage, compute, etc. Takes 7-8 years for storage engine to mature. Worked on DCs and Atlas. Seeing Atlas go from 0-1B in 5 years. 3 people who worked on it. Backup, Automation, Monitoring, etc. DataLake. He knows Josh pretty well!
- How are you being intentional about the culture you've built and continue to build with each new hire?
Answers
- Why MongoDB? People (OSS and interviewers), Lifestyle, Work.
--> People: 1) I owe my relative success in this industry to the mentors who empowered me back when I was a college freshman. I landed an on-campus job that managed Linux packages for the university, and had some of the best coworkers - several of which have been working at MongoDB since they graduated - there must be good reason for that. 2) I've worked with folks at MongoDB through my involvement in hackNY - mostly interns and recruiters and intern managers - and I've become accustomed to the kind and intelligent folks that pervade. 3) The interviewers I've met have been incredibly insightful. Dave (observability and edge) did 2 of my interviews and was a true pleasure to speak to, and I forgot their name, but my Behavioral Interviewer has been at MongoDB for 10-11 years and said that they stick around because every day they work next to some of the most brilliant people in the tech industry.
--> Lifestyle: I've heard from many sources over the years that MongoDB engineering strikes a great balance between work and life. As a recent homeowner in NJ and a first-gen immigrant with dependents, I've been searching for a workplace that would both challenge me intellectually and respect my personal responsibilities - I'm not sure what the optimal balance is yet, but a hybrid work model with 2-3 days per week spent in the office sounds great.
--> Work: The work itself sounds technically challenging and piques my intellectual curiosity. I've spent 5 years at FB and in all that time I've learned a lot about managing an on-prem "cloud", through my work on the platform through which billions of users per second access FB services. But I've realized, especially in this interview loop, that the industry is prevailingly multi-cloud now, and at the same time, I joined FB at a time when most interesting things had already been built. I know Atlas is natively multi-cloud and a relatively recent product, and I'd love the opportunity to dig into its guts, so to speak, and learn more about it, and hopefully deliver some big wins in the process.
- Intro? See below.
- Appreciate: The team. The culture - building things fast. Healthy mix between design and getting stuff done. Flexibility. Slack, zoom, etc.
- Wish was different: Nothing - strong locus of control.

- Maybe one day back on-prem offloading work. They shut off DCs 5 years ago. Blurring between control/data plane on serverless.
- 30 people on SRE. NYC, Dublin, 1 in Madrid, etc. 4 teams: Helix (internal MKS), Observability, Atlas, Special Projects, Compliance. They do affinity, not embedding. Special Projects build special environments: replicas of MKS for FedRamp. AWS gov* are regions. They can touch anything. Government workloads run on Special Projects environments - big source of revenue. Sales cycles are longer. Would run into everyone IRL. This is revenue-generating, growing, etc. Grows much larger within 6mo-2yr. 50/50 code/systems. Middle of moving them to EKS.

Career Journey?

I was a PE on the TPE team within the larger Traffic organization at FB, and I thought about the security, reliability, and scalability of FB's edge platform every day for 5 years. I spent a lot of time tackling super low-level problems, and I took these challenges as opportunities to deliver big wins, grow my team, publicize my team's accomplishments, etc. I became quite adept at translating context between various classes of stakeholders, including ISPs, software engineers, and network contractors.

Prior to that, I received a masters-level CS education from Rutgers University in NJ, with a major focus on Algorithms, AI, and Distributed Systems, AI, and a minor focus on Probability and Astrophysics. I spent 3 years leading student organizations and running HackRU, the first-ever and 2nd-largest student hackathon. And through that involvement, I also became quite involved with a nonprofit in NYC called hackNY which runs a fellowship program that brings together 20-30 brilliant, responsible-minded technologists to live and work together in the heart of NYC, and inducts them into a larger community of hundreds of alumNY. I have been directing this program for the last 2 years.

base 175-200
bonus (perf) with refreshers
equity (tokens) 1 year cliff, monthly
annual reviews

175-200 base
150-200 equity/year - 450-600 total over 3 years
runway? 23m, seed 3.6 - 26m in bank for 2-3 yr runway, only expense is employees, little marketing/pr/office
hasn't yet announced series A because of rebrand in mid-sept
main net is IPO (DEX on coinbase) - 1 yr bcap
bo/peter are night/day - blunt, looking for enthusiasm/cultural fit

CEO: cultural fit + bar raiser, conduct orchestra, show enthusiasm for the industry, greater than the sum of its parts
CTO: build a good team, intellectual curiosity, aspiring exceptional technical contributor / engineering manager, do technical work / manage teams / etc.
VP (NPC, stooge): emotional intelligence, know how to work in large companies, passion for the space, very technically capable
VP (effective): ?

DATADOG

How do I operate?
How does the team operate?

Notes
- Streams - Kafka. On prem or VPC? Have you ever had to dig into its guts for some perf?
- Team. NYC - Growth plan? I'd be on the Cluster Workstream and come up with a 30-60-90d plan.

- Work/life, interesting problems, company direction
- Manage hundreds of kafka clusters and upgrade all brokers
- Go Java Rust Python C

- 9 people, HC 10. Distributed s.t. no 2 people are co-located. 50% in EU, 50% in EST.
- Focus Time and then Collaboration time. It's a Senior-heavy team.
- Next year will be a particularly exciting time to ship. Also will start hiring juniors.
- 2 work streams: Client Work (many kafka clients, one universal client, Go/Rust/-Python/?Java)...
- ...and Cluster Work (automatic scaling, regular partition balancing, no backups, recently 3x RF, EBS on SSDs, cold data to S3, dynamic routing/failover)
- 2-week sprints. oncall rotation: incident response (2d), interrupt handler (daily pages (30-40 per month), responding to people)
- Aim is to get inside all popular VPCs, co-located per client. But they're not multi-cloud. They think about multi-availability zone.

DATADOG 2
- CI/CD, Terraform to cloud providers. provisioning new accounts, orgs. On top is the Kube teams which deploy on their accounts. Cloud Compute. Provisioning IAM, security, etc.
- Gitlab CI/CD, Temporal Workflow Engine. 5 people, across EU and US. 1 sister team of 7: Secrets Management. "Runtime Organization".
- Seniority? Tell a story. That's the thought - they're currently at a pretty healthy distribution. Leonid's team is senior-heavy.
- Growth plan? Is it more clear or more like "there's a lot of work." There's a 30-60-90 day onboarding schedule. They're both managers bc of peer feedback.
- Oncall experience? There is operational load, they're automating themselves out. They're not on any critical path. Mostly they help other teams review PRs and stuff.
- About 50% of requests are self-service, but even those need to be monitored, they're working on a security linter for terraform, etc.
- What's the roadmap? Where do I fit in? "Automating all the things!" Quarterly planning. Setting up a large VPC in Japan. Test DC provisioning up and down.
- I said I didn't like reinventing the wheel, shadow infrastructure, etc.
- Technical scale, global problem space, huge levels of innovation. They're multi-cloud because of customers - it's unique. Amazon, Dropbox, startup, global ad-tech, monitoring.
- Hybrid? Flexible - 2-3 days in the office. Summit in NYC/Paris. We-Work company subscription!!

DATADOG 3
- Senior Director has been there 6 years. Started as EM of metrics aggregation team of 6 people, grew it into 20 people.
- 6-12 months I clearly demonstrate Senior contributions, we'd put promos forward. We do consider promotions on a quarterly cadence.
- Senior contributions: Depth and complexity of RFCs, proposing new ideas and direction, enabling others to execute. Operational excellence, domain expertise, etc.
- Promotion proposal process: Manager writes a packet, scope + impact, leadership, coding, endorsements, etc. Panel review in DDS (4-5 VPs). Considers context of team's tasks.
- What are you excited about? 6 months on Platform Automation. Event Platform built something on top of Kafka, higher level abstractions, and wants to distribute that. Each platform builds their own Control Plane. On the cluster side, Kafka brokers are local instances. They could be EBS, but that would more expensive and less easily snapshot-able. They want to keep local data for ~1h - implementing tiers of storage would dramatically drive down costs and change the way they forklift data every 24h.
- What was your journey to Senior Director? If you were me, how would you get to that goal by 32? They did Deloitte at L3, then Product Digital at L4, quickly moved into a management role. Put yourself in positions where things are growing, and opportunities will come. Personality drove them to management side because they are naturally bossy, and staff+ was less defined.

MONGODB
- Open-source and closed-source mix. CTO just left, was never exciting. No human senses of humor. Everyone else is inspiring.


POLYMER
- This is obviously greenfield. How are you thinking about market strategies here? Are you designing your own "ISP"? Does that even make sense?
	- They are one of the teams contributing to IBC, pushing its adoption, how to make it easier to integrate, p2p (N^2), hub-spoke (N), relay exchange?
	- Computation can be run off-chain. Polymer is that for IBC. No competitors at the TAO layer. The community is abuzz about IBC.
	- Building our own validators, interesting maybe build our own relayer network? Mitigate risk by diversifying income.
- Where do you see the company in 6mo, 1yr, 2yrs? Why are you in this? Is it for the long haul, the philosophy, etc?
	- Launch main net in 6mo. 1yr integrate into Ethereum, launch token. 2yrs get IBC and Polymer broadly adopted. Top 5-10 company in Blockchain space.
	- They need a DevOps person. Automation framework to operate chains on IBC. LayerZero.
	- Financial colonialism. Twitter - random macro things. We are financially colonizing these countries via feds' policies.
	- Want to work on something disruptive, that changes the industry. Building on a 10-20 year timeframe - with variable capacity. Animals, sustainability, mental health.
	- Would probably still be involved somehow in a foundation entity. Founders take things from 0-1 and CEOs take it from 1-100.
	- Psychedelics on mental health, building animal sanctuaries, water conservation, etc.
- What do y'all do for fun? How are you being intentional about the culture you're building with each new hire?
	- Intellectual curiosity. Game nights.

runway? 23m, seed 3.6 - 26m in bank for 2-3 yr runway, only expense is employees, little marketing/pr/office
hasn't yet announced series A because of rebrand in mid-sept
main net is IPO (DEX on coinbase) - 1 yr bcap
bo/peter are night/day - blunt, looking for enthusiasm/cultural fit

EY strategy practice, 2017 whitepaper work, 2020 DAOs. What's your journey? how'd you meet Bo?

I was a PE on the TPE team as part of the larger Traffic org, and I thought about the security, performance, and scalability of FB's edge platform every day for 5 years. I spent a lot of time tackling super low-level problems, and I took these challenges as opportunities to grow my team, publicize my team's accomplishments, etc. I became quite adept at translating esoteric low-level OS context into something that various stakeholders, including ISPs, software engineers, and network contractors, could understand and empathize with.

Besides that, I graduated with a masters-level education from Rutgers University in CS with a focus on Distributed Systems, AI, and Algorithms - also focusing on Math: Probability and Astrophysics. I spent 3 years leading CS student organizations and running HackRU, the first-ever and 2nd-largest student hackathon. I also became quite involved with a nonprofit in NYC called hackNY which runs a fellowship program that brings together 20-30 brilliant and responsible-minded technologists to live and work together in the heart of NYC, and inducts them into a larger community of hundreds of alumNY. I have been directing this program for the last 2 years.

- Plan for operating playbook for a blockchain like ETH. Diff between blockchain and high-level DB.
- VM and Consensus Engine. How does it impact how I operate?
- No outside
- State transition function that applies some logic, R/W applied to KV store. Merkle root -> function -> Merkle root.
- IFTTT run by validators. Constrain problem to PoS model. Safety > liveness. 66% quorum. BFT style PoS priority security + speed of finality, not decentralization.
- ETH: block production (speed) + finality algorithms (security).
- Size +
- Cosmos (150) 1-2sec, ETH (700k) 15m blocks
- N^2 P2P 150 nodes. We operate one node out of 150.
- Playbook: transparent upgrades, security considerations (slashed), DDoS
- Network requirements on our node? We want value out of the network - we want to max our value.
- Speed of reboot. Upgrade approach? 2 nodes running simultaneously. Multiple machines. One live machine.
- Chain is super basic - one set of KV data. Stream between read/write replicas.
- Mechanism for "connecting" your live node to P2P? When node comes online, broadcast identity to network via private key.
- Proposer vs voter. Packets are basic, few message types: Outbound: PROPOSE, VOTE. Inbound: YOU_ARE_PROPOSER, YOU_ARE_VOTER via IP. Secure the key??? Slow ingress/egress.

- Updates: priority tiers: ZD kernel vuln. P0. Kernel/firmware upgrades P1. Destructive.
- On-prem: TPM Quotes for Priv Key distribution? Better than flash USB devices! Split private key into shards.
- Apache Zookeeper so the nodes know which one is the leader. These are all control plane nodes.
- How do you initiate a swap. Some human operator triggers upgrades and prio's them. Try writing them out later.
- Share this IP!! MQ node storing messages, HC node that tracks health of my replicas. Data store. Kube master failover? One node for handling voting logic.
- Human: 1. Schedule upgrades ahead of time, ideally some automation is run to do the work. 2. Validating TPM Quote. 3. Transferring the key and state. 4. Can duplicate outbound messages! 2 validators can run simultaneously. 5. Destructive maintenance after the new validator comes up. Single outbound queue?
- State sync across validators = replay the blocks. Do not start from genesis! Use CE to bootstrap from state snapshots.

- Relayer Hub
- Unidirectional Communication: different types of relayers, offload (different types of) validation work. Light clients are the endpoints of physical connections. IBC default is . Trust relay network - no!!

FD
- It's been 4-5 years since I did docker/kube devops at Oscar and since then I've worked with TW
- I've been doing a lot of interviews but I spent 8 hours yesterday refreshing my mind on Kube/Docker/Nomad so...