Research Scientist / Engineer - Agents

Job Description

Job Description

About ChipStack

Chips are at the center of today's tech-driven world. But how we design them has not changed in decades, while their complexity and specialization have skyrocketed due to increasing performance demands from applications like AI. We want to change that.

Our team is small, technical, and fast-moving. We’ve built and shipped at the intersection of AI, EDA, and systems software, with deep roots at companies like Qualcomm, Nvidia, Google, Meta, and the Allen Institute for AI. We’re backed by top investors including Khosla Ventures, Cerberus, and Clear Ventures, and already deployed with 10+ innovative customers—from Fortune 100s to cutting-edge AI silicon startups.

Role Overview

You’ll design, engineer, and research agentic systems—software that plans, executes, uses tools, and learns. You’ll bridge LLM agent research (prompting, tool use, memory architectures) with the engineering that puts these systems in production for chip design workflows.

Key Responsibilities

Agent Research & Prototyping

Architect and iterate on agentic components : memory, context engineering, tool usage, multi-agent coordination

Experiment with prompting, finetuning, and orchestration to improve agent performance

Infrastructure Development

Build scalable infrastructure for prompt iteration / testing, benchmarking, logging, and evaluation

Assist with pipelines for fine-tuning, automated evaluation, and model deployment.

Collaboration & Integration

Work directly with chip designers and ML infra engineers to embed agentic workflows into chip design pipelines

Partner with product teams to solve core agent challenges (e.g., long-horizon planning, tool integration)

Research & Metrics

Define and evaluate performance metrics for agent systems

Design goal-oriented agentic workflows to enhance task completion

You Should Have

5+ years of combined ML research and / or software engineering experience

Hands-on with LLMs : prompting, finetuning, evaluation

Strong programming skills

Solid foundations in distributed systems, data pipelines

Cloud + on-premise infra experience involving GPUs. Familiarity with Docker

Excellent communication & a bias toward collaborative problem-solving

PhD or MS in ML, CS, or related field, or equivalent research + implementation experience

Ability to build, evaluate, and iterate LLM-based applications

Nice-to-Have

Experience in fine-tuning LLMs specifically for agentic applications

Experience designing agentic systems or multi-agent workflows

Familiarity with reinforcement learning (RLHF / RL-agent contexts)

Background in integrated ML infra and agent evaluation pipelines

Interest or experience in the chip design / EDA domain

What Makes This Role Unique

Play a key role at the intersection of agent research and real-world infra , delivering autonomous workflows for chip design

Influence both the research direction and the productionization of agentic systems

Join a tight-knit founding team, directly impacting high-stakes chip-design automation tasks

Culture

Challenge status quo : We are innovators who can challenge the status quo and push forward our vision of the world.

Strong opinions, loosely held : We are low on ego, but high on collaboration. We are okay to be wrong and are always open to learning.

Ship fast, ship quality : We ruthlessly prioritize what matters. We build a few things, but at lightning speed with high quality.

Proud of our craft : Attention to detail is in our DNA. We take pride in what we build and ensure they exceed the high standards of the semiconductor industry.