← baseten / Product Manager, Developer Experience
tailored_resume_v2 / art_gxLr4Fmx-8s
role
model
anthropic/claude-sonnet-4.6
created
2026-06-11T17:23
↓ Download .docx ↓ Download .pdf PDF requires LibreOffice installed
What changed for baseten
| change | why it matters |
|---|---|
| Summary rewritten to lead with SDK/DevPortal onboarding proof point and RL workbench builder credibility | JD's first requirement is 'built dev tools as an engineer before becoming a PM' and 'shipped SDKs, CLIs, APIs'; summary must immediately establish both engineering and PM credibility |
| Intuit reordered to lead experience section and SDK Starter Kit bullet moved to position 1 | Intuit is the highest-relevance role (score 5); SDK/DevPortal onboarding is the closest analog to Baseten's CLI/SDK ownership and first-run experience mandate |
| Intuit bullet 1 reframed to explicitly connect SDK Starter Kit work to Baseten's first-run experience problem | JD asks PM to own 'first-run experience — onboarding, model discovery and deploy'; making the parallel explicit strengthens perceived fit |
| Intuit Drift Detection bullet reframed as 'closing the deployment lifecycle loop' | JD explicitly lists 'deployment lifecycle and safe releases' as a core ownership area; connecting the drift detection work to that language is accurate and relevant |
| Streamio reframed to lead with OpenClaw agent-first orchestration and MCP SDK bullet | JD states 'treat the coding agent as a first-class user, not a feature' and 'agent-driven ways developers build'; MCP SDK work is the strongest proof of agent-first design conviction |
| Fintellect condensed to 2 bullets focused on multi-provider LLM routing and production model serving | JD bonus asks for 'experience with ML infrastructure, developer tools, or API products, you've deployed and served models yourself'; Fintellect's RAG/LLM routing is the proof |
| RL Workbench moved to lead the projects section | Baseten just launched Loops (Training SDK for Frontier RL workloads); RL Workbench benchmarking TRL/VeRL/OpenRLHF/NeMo RL is the single strongest signal of ML infrastructure depth |
| RL Workbench bullet 1 explicitly references Baseten Loops alignment | Directly connecting candidate's project to Baseten's May 2026 product launch demonstrates company research and product intuition |
| IBM role condensed to 1 bullet reframed as 'shipped software as an engineer before transitioning to PM' | JD hard requirement is 'built dev tools as an engineer before you became a PM'; IBM establishes this credential without consuming space |
| Bank of America role removed from experience section | Zero relevance to developer experience or ML infrastructure; space better used for technical proof points |
| Deep Learning Education Platform project removed | Lower relevance to Baseten's DevEx PM role than RL Workbench, aeval, BRAIN, and AutoEval; space optimization for 2-page target |
| Lawrence Berkeley National Laboratory entry removed from projects | BRAIN project already captures the NeurIPS/protein structure work; Lawrence Berkeley entry is redundant and lower impact |
| Section order: Experience → Projects → Education → Teaching → Additional | Baseten values technical builders; projects section (RL Workbench, aeval) provides critical ML infrastructure credibility that supports the experience section claims |
JD analysis (20 key phrases)
Key phrases: developer experienceCLI and SDKsfirst-run experiencedeployment lifecyclemodel servingagent-first designself-serve deployproduction deploymentdeveloper-facingonboardingmodel discoveryprogressive deliverymulti-model compositionvoice of the developerplatform surfacesSDKAPI-firstinferencedeveloper toolingship great product
Hard requirements:
- Built dev tools as an engineer before becoming a PM
- Shipped SDKs, CLIs, APIs, or platform surfaces hands-on
- 5+ years building developer-facing or technical products
- Technical enough to earn engineers' trust
- Strong opinions on CLI, API-first, and agent-first design
- Customer-obsessed researcher who lets reality drive roadmap
Preferred qualifications:
- Experience with ML infrastructure or API products
- Deployed and served models yourself
- Experience with developer tools products
Per-role mapping (11 roles scored)
| role | score | reframe angle | JD phrases that map |
|---|---|---|---|
| Intuit — Staff PM, Developer Frameworks & Platform Infrastructure | 5/5 | SDK/CLI developer tooling, self-serve onboarding, platform infrastructure at scale | CLI and SDKs, first-run experience, onboarding, developer-facing, platform surfaces, self-serve deploy, deployment lifecycle |
| Streamio AI — Founder & CEO | 4/5 | Agent-first platform design, MCP SDK tooling, multi-agent orchestration, 0-to-1 product shipping | agent-first design, CLI and SDKs, multi-model composition, ship great product, platform surfaces |
| Fintellect AI — Founder & CEO | 3/5 | Production LLM routing and model serving, multi-provider orchestration | model serving, inference, API-first, production deployment |
| Splunk — Senior PM, Search Orchestration | 3/5 | Microservices platform PM, performance optimization, developer-facing search APIs | API-first, deployment lifecycle, developer-facing, ship great product |
| Kaiser Permanente — SOA Technical PM | 2/5 | Platform infrastructure at enterprise scale | platform surfaces, deployment lifecycle |
| IBM — Software Engineer, Business Intelligence | 2/5 | Engineering credibility — shipped software before becoming a PM | technical enough to earn engineers' trust, built dev tools as an engineer |
| Bank of America Merrill Lynch — Tech MBA Summer Associate | 1/5 | Quantitative analysis background | — |
| RL Workbench — Post-Training RL Platform | 5/5 | Hands-on ML infrastructure builder with RL post-training and model serving depth | model serving, inference, deployed and served models yourself, ML infrastructure |
| aeval — AI Model Evaluation Platform | 4/5 | Production model evaluation infrastructure, CI/CD for ML | deployment lifecycle, progressive delivery, model serving, inference |
| AutoEval — Automated Visual Evaluation for Robot Model Training | 3/5 | ML model evaluation tooling, multimodal inference | model serving, inference, developer tooling |
| BRAIN — Protein Structure Prediction ML Platform | 3/5 | End-to-end ML platform builder with model serving and experiment tracking | model serving, inference, deployed and served models yourself, ML infrastructure |
Tailored summary
Technical PM with 12+ years building developer-facing platforms, SDKs, and AI infrastructure — from shipping SDK Starter Kits and a self-serve DevPortal that cut developer onboarding from weeks to minutes (Intuit, 675M+ engagements, 50K TPS) to hand-building RL post-training workbenches that benchmark GRPO/DPO across TRL, VeRL, OpenRLHF, and NeMo RL today. Deep conviction on CLI, API-first, and agent-first design — I've deployed and served models myself, built multi-agent orchestration frameworks, and treat the coding agent as a first-class user. NeurIPS published. UC Berkeley BS Computational Engineering.