Build Agentic AI system for analog chip design

• Top VC backed | DoD contractor (Defense Innovation Unit)

• We're hiring! Drop me a message for more info

Key words: Agentic AI (Knowledge, Context, Memory, Infra, Eval), Silicon EDA (Analog, P&R, PDK, Rust), GUI (WebGL, React), Distributed System

Push EDA tool PPA frontiers with AI

• Building deep learning opt. engine in Innovus (data pipeline, deploy, debug/eval infra)

• Core power opt. engine dev

• Customer engagement (e.g. Nvidia, Apple, etc.; ~ 3% power gains in 2nm blocks)

• LLM digital EDA backend (debug tools, codegen, copilot)

Key words: GNN, RL, LLM, RAG, Digital Design; PyTorch, PyGeometric, Python, C++

Research Thesis: DL, RL, Optimization, GNN, Applied ML (Chip Design, combinatorial problems, time series model, etc.); Lab: Graphics Group

Research Project: DARPA IDEA, in collaboration with Cadence & Nvidia

Advisor: Levent Burak Kara, Barnabas Poczos

Audio ML team (got full-time AS II offer)

Project: RL Neural Architecture Search (NAS) for Speech Enhancement model on Neural Network Accelerator (NNA)

Key Words: NAS, model compression/quantization, ML accelerator, model-platform co-design, Audio ML, RL

NVIDIA Research (ASIC's team)

Project: RL, GNN, Transformer for Placement and Route of Advanced Node Standard Cell

Key Words: RL, chip design, GNN, Transformer, combinatorial optimization, imitation learning