Intern - Machine Learning Engineer

What you'll do

Fine-tune, evaluate, and deploy LLMs and foundation models for specialized chip engineering tasks.

Build scalable inference systems, agentic workflows, and RAG pipelines to support intelligent, interactive design tools.

Optimize model performance through batching, caching, quantization, and streaming to achieve sub-second latency.

Collaborate closely with AI scientists to operationalize experimental techniques and deliver robust, production-grade systems.

Contribute to architecture decisions around scalable, reliable ML infrastructure in a cloud-native environment.

What You Bring

Hands-on experience working with LLMs, including prompt tuning, fine-tuning, and evaluation.

Deep expertise in Python, PyTorch, and the surrounding ML tooling ecosystem.

Familiarity with orchestration frameworks and agentic workflows (e.g., LangChain, LlamaIndex).

Strong understanding of vector search, semantic embeddings, and RAG systems.

Solid grasp of scalable ML architecture and the principles behind high-availability inference systems.

Bonus Points

Experience working directly with ML researchers or in applied research settings.

Startup experience or contributions to high-velocity, cross-functional teams.

Open-source contributions to ML tooling or infrastructure projects.

Experience deploying models in cloud environments (e.g. AWS SageMaker, Bedrock, vLLM, or SGLang)

What it's like here

We’re a fast-moving AI startup with a collaborative, high-trust culture. You’ll work side-by-side with top engineers and scientists, translating cutting-edge research into real product impact. We value curiosity, speed, and precision—engineering systems that are elegant, pragmatic, and built to last. If you’re passionate about bringing frontier Physical AI to the real world, you’ll feel right at home.

Logistics

This position is available in Mumbai, India. We believe a mindmelt between engineers and scientists leads to our professional growth and to great products; we have a hybrid schedule.