Vivek Bhadauria

Tinkering with systems | Backend & Infra Engineer | Kubernetes, eBPF, distributed systems

[Vivek Bhadauria @viveksb007]

Gave a talk at Amazon Kubernetes Summit on "Life of a DNS request in K8s and DNS Network Policies".

Walked through the full journey of a DNS request inside a k8s cluster - from Application pod to VPC Network to CoreDNS and back.

Talk is mostly extension of the my deep dive post - https://t.co/0qrfA390Pn

[Vivek Bhadauria @viveksb007]

@valyala @janetacarr "software engineering" is supposedly solved. now we are in "software factory" era

[Hamel Husain @HamelHusain]

This is the most valuable free resource we've created on AI Evals 🎉 (not exaggerating!)

I organized our public materials into this guide. It allows you to find answers to your eval problems w/o searching aimlessly.

Humans: pick the row that sounds like you.
Agents: point it at the post

The material draw on 60+ hours of office hours from our Evals course, where @sh_reya and I have taught 5k engineers and PMs Evals.

We add new material often, with 15 FAQs added in the last two weeks. Recent additions are marked with a "New" badge. You can find all of these and more here:

https://t.co/yVp03nPmbY

[Vivek Bhadauria @viveksb007]

@trq212 I also use plan mode as a discussion phase to come to consensus other model goes ahead and starts doing things many times (i have to explicitly mention, don't do anything let's discuss first)

[Vivek Bhadauria @viveksb007]

@shrey_sancheti kagent maintains a session on postgres which tell what all happened as part of that execution. they have a default UI for HITL. slack is just another client surface for similar work
check out - https://t.co/SRaTqisOB9

[Vivek Bhadauria @viveksb007]

slack-kagent integration for tool-approval is looking good

[Vivek Bhadauria @viveksb007]

@kalamazooooo Is that @deepigoyal ?

[Vivek Bhadauria @viveksb007]

@jordancodeadv @ChShersh TIL - i was also loosely thinking they are same things

[Vivek Bhadauria @viveksb007]

what are your thoughts on Managed Agent Platforms like Claude Managed Agents?
i assume there will be a mix of cx
- some choosing fully managed and give away flexibility of trying out different things like other models and harness combinations.
- others taking operational aspects and running platform on their own

[Vivek Bhadauria @viveksb007]

Data is definitely one of the most important pillar for Advanced AI development (similar to compute). Crazy growth trajectory 🚀🚀

[Alex Ratner @ajratner]

I'm excited to announce @SnorkelAI's $350M Series E at $3.5B, led by @insightpartners and @S32_VC.

We've grown 18x+ in the last 12 months since launching our Data-as-a-Service offering, passing $375M ARRR this week.

As AI advances to superhuman capabilities, AI data & environment development must advance with it - and basic staffing and crowdsourcing approaches are not enough.

AI progress now requires deep research and technology work that combines human expertise with specialized AI in compounding ways. @SnorkelAI is building the RSI data engine and frontier data lab for this next phase.

We're honored to have the support of existing investors Addition, @lightspeedvp, @GreylockVC, @GVteam, P7, Factory, @WellsFargo, Walden Catalyst Ventures, and new investors @ThirdPointLLC, @MarchCPs, @BlumbergCapital, @AllegisCapital, @Frontlinevc, and @standard_vc.

–

@SnorkelAI started as a research project a decade ago at @StanfordAILab.

Our thesis was simple: AI progress would become increasingly data-centric – and therefore data development should be studied as a true research and technology problem, not just a staffing and crowdsourcing one.

Today, as AI capabilities verge on superhuman, building the data and environments to safely measure and train AI is becoming too hard for even the smartest human experts to do alone.

Only humans and AI agents, collaborating together in compounding ways, can meet the accelerating needs of the frontier, and keep humans in the driver’s seat of AI progress for decades to come.

At @SnorkelAI, we are building the data lab to define the shape of this new “Data 2.0” frontier, and the new paradigms of human-computer interaction needed to advance it.

Our key focus is building the RSI engine for data, where specialized AI models accelerate and improve human expert output, and in turn, scaled human supervision is used to continuously evaluate and improve these models – creating a powerful compounding loop to keep pace with an accelerating RSI frontier.

With this round of funding, we are also doubling down on our commitments to support data development for open benchmarking and evaluation (more news here soon!); an increasingly diverse ecosystem of general and specialized intelligence; and a path to safe, well-aligned AI built on robust training and evaluation data.

Data development will guide and drive the next stages of AI – and must do so in a human-centric, AI accelerated, open, diverse, and safe way. We are excited to support this mission in the next decade of research ahead at @SnorkelAI.

More thoughts here: https://t.co/Dzk6olqmAc

[Vivek Bhadauria @viveksb007]

@rakyll Then we have sometime before that happens 😅

[Vivek Bhadauria @viveksb007]

Learn DNS basic from this cool challenge.

Check out https://t.co/x6aGJENeiu which walks through how DNS works in k8s pods

[Ivan Velichko @iximiuz]

DNS 101: Resolving public, internal, and local-only hostnames 🧙

Most of the time, when you use curl or ssh, you point it to a hostname. When a network request uses a name instead of an IP address, it must resolve the name first.

Learn how it happens: https://t.co/oyercC5XXR

[Vivek Bhadauria @viveksb007]

@AbhiCodes15 Probably better to answer who is not getting fired
- a developer who ships good and maintainable stuff (using AI to increase shipping speed and automate mundane tasks)

[Vivek Bhadauria @viveksb007]

@valyala @neerajjj6785 Did you migrated from k8s to directly managing VMs on AWS?

[Vivek Bhadauria @viveksb007]

@aarondfrancis @ashebytes thanks for writing it. i read it around a year back and that inspired me to share more of the work that I do

[Vivek Bhadauria @viveksb007]

@awsdevelopers @grok answer this tweet

[Vivek Bhadauria @viveksb007]

@codewith55 Whatever you are comfortable with and able to read and understand the code

[Vivek Bhadauria @viveksb007]

@thdxr OpenCode Rust

[Vivek Bhadauria @viveksb007]

@rawkode Pretty cool!!
Curious about which models you tried with and how much time/tokens they took? Just for comparison across the models fixing the cluster

[Vivek Bhadauria @viveksb007]

@thegeeknarrator @bookingcom Had same experience recently, they outsource they cx support to GoTo support or something and its pretty poor cx handing coupled with poor UI/UX too (whenever I went to some other app while waiting, my existing waiting support case was lost, then I had to start again)

[Vivek Bhadauria @viveksb007]

@MaxBrodeurUrbas Very cool!!
Do you folks host the open weight model or route to some other open model provider like Fireworks, Baseten?
(Asking from data POV)