← supabase / Product Manager - AI
tailored_resume_v2 / art_lIN9eGU179c
↓ Download .docx ↓ Download .pdf PDF requires LibreOffice installed
What changed for supabase
| change | why it matters |
|---|---|
| Summary rewritten to lead with MCP server and agent-native stack credentials | JD's first and most specific requirement is ownership of the MCP server and agent skills — leading with this immediately signals fit |
| StreamIO bullet 1 reframed to foreground MCP server and AI coding assistant integration | JD explicitly states 'You own the MCP server' — this is the single most differentiating proof point on the resume |
| OpenClaw bullet reframed around 'tool boundaries and defaults' language | JD emphasizes 'defaults and abstractions' as the core PM judgment call for agent-facing products |
| aeval project moved to lead the Projects section | JD states 'evals that tell you whether any of it is working' — aeval is the most direct proof of eval infrastructure ownership |
| Fintellect condensed to 2 bullets emphasizing multi-agent reliability patterns | Relevant for agent orchestration credibility but lower priority than StreamIO/Intuit; space optimization |
| Intuit bullets reordered to lead with scale metrics (675M+ engagements, 50K TPS) | JD targets candidates who have shipped at scale; enterprise metrics establish credibility before SDK/DevPortal details |
| Splunk condensed to 2 bullets | Relevant for microservices/backend PM credibility but lower priority than founder and Intuit roles for this JD |
| Kaiser condensed to 1 bullet | Weakest fit role; retained for career continuity and platform infrastructure signal but space-optimized |
| Bank of America role removed from experience section | Lowest relevance to developer tools / agent PM role; space reclaimed for higher-signal content |
| Summary embeds 'agent-native', 'MCP server', 'agent skills', 'evals', and 'developer activation' language | These are exact key phrases from the JD; embedding them accurately signals language alignment without fabrication |
| aeval project bullets explicitly note applicability to 'agent skill eval infrastructure' | JD asks PM to define how launches are measured via AI eval scores — connecting the project to the job requirement directly |
JD analysis (20 key phrases)
Key phrases: MCP serveragent skillsevalsagent-friendlyagent-nativeAI coding toolsdeveloper activationtask completionAI eval scoresbreaking changesdefaults and abstractionsship fastdeveloper toolsinfrastructureasync by defaultroadmap prioritizationindie developers to enterprise teamsagents writing migrationsopen-source-firstbuild in public
Hard requirements:
- 7+ years PM experience on developer tools OR ex-founder with strong product instincts
- Technical depth: read architecture docs, follow design discussions
- Owns MCP server, agent skills, and evals
- Clear POV on agent-friendly infrastructure
- Daily use of AI coding tools and agent harnesses
- Bias toward speed — ship and learn
- Async-first work style
- Define success metrics before engineering starts
Preferred qualifications:
- Shipped something that tested opinions on agent-native primitives
- Tracks Anthropic, OpenAI, and AI community weekly
- Uses AI tools for research, drafting, synthesizing feedback
- Experience with CLI, SDK, and dashboard tooling
- Cross-functional alignment across engineering, design, leadership
Per-role mapping (11 roles scored)
| role | score | reframe angle | JD phrases that map |
|---|---|---|---|
| Streamio AI — Founder & CEO | 5/5 | Agent-native product builder who has shipped MCP server tooling, multi-agent orchestration, and AI coding assistant integrations in production | MCP server, agent skills, AI coding tools, agent-native, ship fast, defaults and abstractions, developer tools |
| Fintellect AI — Founder & CEO | 3/5 | Multi-agent orchestration and LLM infrastructure experience; customer discovery discipline | agent skills, AI coding tools, customer discovery |
| Intuit — Staff PM | 5/5 | Developer platform PM who shipped SDKs, DevPortals, and infrastructure tooling at massive scale — directly analogous to Supabase's CLI/SDK/dashboard surface area | developer tools, developer activation, ship fast, infrastructure, breaking changes, roadmap prioritization, cross-functional alignment |
| Splunk — Senior PM | 3/5 | Backend microservices PM with strong prioritization discipline and developer-facing API ownership | infrastructure, roadmap prioritization, developer tools |
| Kaiser Permanente — SOA Technical PM | 2/5 | Platform infrastructure at enterprise scale | infrastructure |
| IBM — Software Engineer | 1/5 | Engineering foundation underpinning technical PM credibility | — |
| Bank of America — Tech MBA Associate | 1/5 | Quantitative decision-making | — |
| RL Workbench | 4/5 | Directly maps to 'evals that tell you whether any of it is working' — built the exact class of tooling Supabase needs for agent skill evaluation | evals, AI eval scores, agent skills |
| aeval — AI Model Evaluation Platform | 5/5 | Directly analogous to the eval infrastructure Supabase needs to measure agent skill reliability and task completion | evals, AI eval scores, task completion, developer activation |
| AutoEval | 3/5 | Eval automation and speed — maps to Supabase's need for fast feedback loops on agent behavior | evals, ship fast |
| BRAIN — Protein Structure Prediction | 2/5 | Deep ML research credibility | — |
Tailored summary
Technical Product Manager with 12+ years shipping developer-facing platforms and AI tooling at scale — including production MCP server integrations, multi-agent orchestration frameworks, and AI eval infrastructure built from scratch. Shipped the full agent-native stack: MCP SDK, subagent delegation, and automated evals that benchmark model behavior end-to-end. Scaled developer platform infrastructure to 675M+ engagements at Intuit (50K TPS, 1.5M concurrent connections) and cut developer onboarding from weeks to minutes. Ex-founder, NeurIPS published researcher, and daily user of AI coding tools with strong opinions on the primitives developers need to build agent-native products.