Looking for a job

sunil kumar

@sunilkumar37

Message

I build production GenAI, agentic, and distributed ML systems for engineering workflows.

India

Message

What I'm looking for

I'm looking to build production GenAI, multi-agent, RAG, and distributed ML systems where I can own work from raw data and architecture through deployment, evaluation, and safe enterprise operations.

At DEP India – Meshwork AI, I build GNN training and prediction systems for automotive FEM surrogate problems, achieving 90%+ accuracy while cutting tensor-creation time by 40%.

I re-architected large-graph training with PyTorch FSDP, DDP, and lazy DataLoaders to eliminate OOM errors, reduce GPU memory by 35%, and cut end-to-end processing time by 40%. I also delivered CNN and GCN surrogates, Transformer forecasts, and FastAPI/Docker deployments that reduced automotive test cycles by 30%.

I'm currently building enterprise autonomous coding and multimodal RAG systems with plan-DAG execution, hybrid retrieval, sandboxed testing, self-review, guardrails, and auditable multi-agent orchestration.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

DEP India – Meshwork AI India

Jul 2024 - Present (2 years 2 months)

Built GNN training/prediction systems for automotive surrogates, cutting tensor creation time by 40% and achieving 90%+ accuracy. Delivered CNN scalar surrogates and Transformer forecasts, reducing client test cycles by 30%.