← nvidia / Senior Product Manager, AI Frameworks
tailored_resume_v2 / art_tvYwYse1V1U
role
model
anthropic/claude-sonnet-4.6
created
2026-05-20T22:03
↓ Download .docx ↓ Download .pdf PDF requires LibreOffice installed
What changed for nvidia
| change | why it matters |
|---|---|
| Summary rewritten to lead with 'AI frameworks, developer-facing platforms, and post-training infrastructure at scale' | JD's first sentence is about building best-in-class AI frameworks; mirroring this language immediately signals fit |
| Summary embeds 'training/post-training landscape', 'E2E ML lifecycle', 'research-to-production', 'large-scale distributed systems', 'OSS-style developer product', and 'GPU-aware framework building' | These are the 5 most-repeated JD signal phrases across hard requirements and preferred qualifications |
| Intuit moved to lead Experience section (was already first but reframed lead bullet to scale metrics) | 675M engagements / 50K TPS is the strongest proof of large-scale distributed systems PM experience — JD hard requirement |
| Intuit SDK bullet reframed to include 'OSS-style developer product delivery' | JD preferred qual explicitly calls out 'Open Source & GitHub-first developer products with deep customer interactions' |
| RL Workbench moved to lead Projects section | Post-training RL platform benchmarking TRL/VeRL/OpenRLHF/NeMo RL with GPU Docker passthrough is the single most relevant project for a role owning AI framework post-training |
| RL Workbench second bullet ends with 'directly applicable to optimizing AI frameworks on NVIDIA GPUs' | Explicitly bridges candidate's GPU-aware benchmarking work to NVIDIA's platform context without fabricating experience |
| Splunk performance optimization bullet reframed to note 'directly analogous to GPU performance profiling and optimization' | JD preferred qual asks for GPU performance profiling; 10x query perf improvement is the closest honest analog |
| BRAIN project repositioned as third project (after RL Workbench and aeval) | NeurIPS publication and 8B-parameter PyTorch platform demonstrate deep learning research credibility required for frontier model builder advocacy |
| Fintellect condensed to 2 bullets; Bank of America reduced to 0 standalone bullets (merged into IBM role context) | Bank of America has minimal relevance to AI frameworks PM role; space reclaimed for higher-signal content |
| Bank of America role removed from sections output | Role has relevance score 1 and no JD phrase mapping; space better used for RL Workbench detail. IBM retained with 1 bullet per hard rule. |
| Deep Learning Education Platform project removed from Projects section | Lower relevance than RL Workbench, aeval, BRAIN, and AutoEval for this specific role; page budget constraint |
| Lawrence Berkeley National Laboratory entry removed from Projects section | BRAIN NeurIPS bullet already captures the research credential; LBNL entry adds length without incremental signal for this role |
JD analysis (20 key phrases)
Key phrases: AI frameworkspost-trainingpre-training/inferenceRecSys and Generative Recommenderfrontier model builderstraining/post-training landscaperesearch-to-productionE2E ML lifecycleproduct strategy, roadmaps, and go-to-market plansopen sourceGPU use caseslarge scale distributed systemsperformance optimizationdeep learningOSS frameworksNVIDIA PlatformHW/SW co-designGPU architecturedeveloper productscustomer interactions
Hard requirements:
- Experience with design and scaling of training/inference/post-training/optimization software (Torchtitan, FSDP)
- Demonstrable knowledge of GenAI or ML concepts: model training, performance optimization, inference, software development
- Experience with large-scale distributed systems
- BS or MS in CS, CE, or equivalent
- 10+ years technical product management at a technology company
- Strong communication and interpersonal skills
Preferred qualifications:
- Experience leading GR systems — GEM, TIGER
- Working on Open Source & GitHub-first developer products with deep customer interactions
- Knowledge of GPU architecture, HW/SW co-design, and performance profiling
Per-role mapping (11 roles scored)
| role | score | reframe angle | JD phrases that map |
|---|---|---|---|
| Intuit — Staff Product Manager, Developer Frameworks & Platform Infrastructure | 5/5 | Large-scale AI/ML platform infrastructure PM with developer framework ownership and research-to-production pipeline expertise | developer products, large scale distributed systems, product strategy, roadmaps, E2E ML lifecycle, performance optimization, OSS frameworks, customer interactions |
| RL Workbench — Post-Training RL Platform | 5/5 | Hands-on post-training framework builder with multi-framework benchmarking and GPU performance profiling | post-training, pre-training/inference, AI frameworks, performance optimization, GPU use cases, training/post-training landscape, OSS frameworks, deep learning |
| aeval — AI Model Evaluation Platform | 4/5 | ML evaluation infrastructure for research-to-production quality gates | research-to-production, E2E ML lifecycle, performance optimization, deep learning |
| Streamio AI — Founder & CEO | 3/5 | Founder-built AI developer platform with multi-agent orchestration and go-to-market execution | product strategy, roadmaps, and go-to-market plans, AI frameworks, developer products, OSS frameworks |
| Fintellect AI — Founder & CEO | 2/5 | Multi-LLM orchestration and GenAI product delivery | GenAI, product strategy, go-to-market plans, deep learning |
| Splunk — Senior Product Manager, Search Orchestration | 3/5 | Distributed systems PM with measurable performance optimization track record | large scale distributed systems, performance optimization, product strategy, roadmaps |
| Kaiser Permanente — SOA Technical Product Manager | 2/5 | Enterprise-scale platform infrastructure PM | large scale distributed systems, performance optimization |
| BRAIN — Protein Structure Prediction ML Platform | 4/5 | NeurIPS-published ML researcher with production deep learning platform spanning 19M-fold parameter scale | deep learning, E2E ML lifecycle, model training, performance optimization, research-to-production |
| IBM — Software Engineer, Business Intelligence Products | 1/5 | Enterprise software engineering foundation | — |
| Bank of America Merrill Lynch — Tech MBA Summer Associate | 1/5 | Quantitative analytical foundation | — |
| AutoEval — Automated Visual Evaluation for Robot Model Training | 3/5 | AI-accelerated model evaluation infrastructure for research-to-production | research-to-production, E2E ML lifecycle, performance optimization |
Tailored summary
Technical Product Manager with 12+ years building AI frameworks, developer-facing platforms, and post-training infrastructure at scale — from hand-coding BPTT in C++ to benchmarking GRPO/DPO across TRL, VeRL, OpenRLHF, and NeMo RL on GPU today. NeurIPS-published ML researcher with deep knowledge of the training/post-training landscape and E2E ML lifecycle. Scaled large-scale distributed systems to 675M+ engagements and 50K TPS at Intuit; shipped developer SDKs, OSS-style DevPortals, and research-to-production tooling adopted across 30+ product SKUs. Rare blend of product strategy, technical depth, and hands-on GPU-aware framework building.