← thinkingmachines / Research Product Manager
tailored_resume_v2 / art_gIPmud091wE
role
model
anthropic/claude-sonnet-4.6
created
2026-05-20T03:30
↓ Download .docx ↓ Download .pdf PDF requires LibreOffice installed
What changed for thinkingmachines
| change | why it matters |
|---|---|
| Projects section moved to lead position before Professional Experience | JD explicitly values past AI publications and research lab experience; RL Workbench and aeval directly demonstrate post-training and evals expertise — the two most cited preferred qualifications |
| Summary rewritten to lead with NeurIPS publication and post-training/evals research credentials | JD's preferred qualifications prioritize past AI publications and frontier research lab experience; leading with research identity maximizes fit signal |
| RL Workbench moved to lead the projects section | Post-training (GRPO/DPO/PPO) and framework benchmarking (TRL, VeRL, OpenRLHF, NeMo RL) are the most direct match to JD's post-training and evals focus |
| aeval positioned second in projects | Evals is explicitly called out as a preferred publication/contribution domain in the JD; statistical rigor and safety testing signal scientific excellence |
| AutoEval added with multimodal AI framing | JD lists multimodality as a preferred contribution domain; AutoEval demonstrates applied multimodal AI evaluation bridging research and production |
| Intuit reordered to lead Professional Experience and reframed around cross-functional program management and scale metrics | 675M+ engagements and cross-functional alignment across 30+ SKUs directly maps to JD's requirement to drive large-scale research products and programs |
| Intuit bullet 3 reframed to highlight 'abstract at high level and get into the weeds' | JD uses this exact phrase as a success signal; Service Language Assessment across 9 languages presented to CTO is a genuine match |
| Streamio AI reframed as 'AI Research Platform' and OpenClaw bullet led | Multi-agent orchestration and MCP SDK integration are the most research-relevant signals from this role; JD values bridging frontier research and real-world applications |
| Splunk bullets reframed around 'translating technical ideas into actionable milestones' and 'identifying bottlenecks' | These are verbatim JD phrases that accurately describe the Scheduler Service delivery and performance optimization work |
| Kaiser Permanente reframed around 'compute and resource roadmaps' and 'identifying bottlenecks' | JD explicitly calls out creating and maintaining compute and resource roadmaps; capacity planning work is a genuine match |
| Deep Learning Education Platform project removed | Lower relevance to research PM role; space optimization to keep resume at 2 pages with higher-signal content |
| Lawrence Berkeley National Laboratory retained as standalone project entry | Academic research lab experience is a preferred qualification; computational biology work under a named PI signals research credibility |
JD analysis (18 key phrases)
Key phrases: collaborative general intelligencefrontier AI researchresearch program managementcross-functional effortscompute and resource roadmapsmodel developmentpost-trainingevalsscientific excellencetranslate technical ideas into actionable plansbridge frontier research and real-world applicationsfast-moving, ambiguous environmentsdeeply technical discussionssynthesize and communicate progressinfrastructure and appliedmilestones and keeping teams alignedproduction systems and product roadmapsdata campaigns
Hard requirements:
- Degree in CS, AI, mathematics, physics, engineering, or similar
- Experience in research program management or product management
Preferred qualifications:
- Masters or PhD in CS, AI, mathematics, physics, engineering, or similar
- Past experience at frontier or academic research lab contributing to AI research
- Past publications relevant to AI or frontier models (evals, multimodality, post-training, pre-training, data)
- Ability to learn new technical domains quickly
- Strong technical communication, written and verbal
Per-role mapping (11 roles scored)
| role | score | reframe angle | JD phrases that map |
|---|---|---|---|
| Streamio AI — Founder & CEO | 3/5 | Frontier AI product builder — multi-agent orchestration, LLM integration, 0-to-1 execution in ambiguous environments | fast-moving, ambiguous environments, translate technical ideas into actionable plans, bridge frontier research and real-world applications |
| Fintellect AI — Founder & CEO | 2/5 | Applied AI product with multi-provider LLM orchestration and structured output validation | bridge frontier research and real-world applications, model development |
| Intuit — Staff Product Manager | 3/5 | Large-scale cross-functional program management, infrastructure roadmapping, and developer platform at enterprise scale | cross-functional efforts, compute and resource roadmaps, keeping teams aligned, milestones, synthesize and communicate progress |
| Splunk — Senior Product Manager | 3/5 | Technical program delivery in fast-moving infrastructure environment with measurable performance outcomes | translate technical ideas into actionable plans, milestones, deeply technical discussions, bottlenecks |
| Kaiser Permanente — SOA Technical PM | 2/5 | Infrastructure-scale program management with capacity and resource planning | compute and resource roadmaps, identifying bottlenecks |
| IBM — Software Engineer | 1/5 | Technical foundation in enterprise software | — |
| Bank of America — Tech MBA Associate | 1/5 | Quantitative analysis background | — |
| RL Workbench | 5/5 | Hands-on post-training research platform directly aligned with frontier model development | post-training, evals, frontier AI research, model development, scientific excellence |
| aeval — AI Model Evaluation Platform | 5/5 | Frontier-aligned evals platform with statistical rigor and safety testing | evals, frontier AI research, scientific excellence, post-training |
| BRAIN — Protein Structure Prediction | 4/5 | Published AI researcher with deep ML foundations from academic lab to frontier scale | past publications relevant to AI, frontier AI research, model development, scientific excellence |
| AutoEval | 4/5 | Applied evals research bridging frontier multimodal AI and production robotics pipelines | evals, multimodality, bridge frontier research and real-world applications |
Tailored summary
Research-oriented Technical Product Manager with 12+ years driving complex AI/ML programs from frontier research to production — NeurIPS published researcher (protein structure prediction, 2014) with hands-on post-training and evals platforms benchmarking GRPO/DPO across TRL, VeRL, OpenRLHF, and NeMo RL. Proven ability to translate deeply technical ideas into actionable plans and keep cross-functional teams aligned at scale (675M+ engagements, Intuit). Thrives in fast-moving, ambiguous environments bridging frontier AI research and real-world applications. MS Software Management, Carnegie Mellon; MBA, Tepper; BS Computational Engineering, UC Berkeley.