Location: Bengaluru, KA, India
Job Type: Full-time
Salary: $30K - $60K
Equity: 0.10% - 2.00%
Experience: 3+ years
Skills: Python, Torch/PyTorch, Reinforcement learning (RL), AI Agents
Company Description
Atrisa (YC P26) builds AI agents for analog circuit design. We are building AI that takes a circuit from specification to layout, working with the simulators and design tools engineers use today.
Role Description
Full-time, hybrid in Bangalore.
Train and fine-tune models for analog circuit design.
Build reinforcement learning pipelines, simulation environments, reward functions, and evaluations grounded in circuit performance.
Build agents that interact with circuit simulators and design environments, from setting up a run to interpreting results and choosing the next step.
Build observability into agent runs. Detect and prevent failures, and design recovery paths when tools, simulations, or model decisions go wrong.
Work with analog engineers to distill their design intuition into training data, models, and agent harnesses.
Qualifications
At least 3 years of relevant AI / ML engineering experience, with hands-on model training or fine-tuning.
Strong Python skills and experience with PyTorch or an equivalent framework.
Experience building agents that call tools, manage state, and carry out multi-step tasks.
Experience designing evaluations and debugging failures across models, data, and software.
Familiarity with reinforcement learning, reward design, or training environments is a plus.
Experience with analog circuits, EDA tools, SPICE, or scientific computing is a plus.
Summary
Train and fine-tune models for analog circuit design, build RL pipelines and evaluations, and create agents that interact with circuit simulators. Requires at least 3 years of AI/ML engineering, strong Python, PyTorch or equivalent, and experience building tool-using agents and debugging model failures.
Responsibilities
Train and fine-tune models for analog circuit design; build RL pipelines, simulation environments, reward functions, evaluations; build agents interacting with circuit simulators; build observability, failure detection, and recovery; work with analog engineers to distill design intuition into data, models, and harnesses
Qualifications
At least 3 years relevant AI/ML engineering with hands-on model training or fine-tuning; strong Python; PyTorch or equivalent; experience building tool-calling agents and multi-step tasks; designing evaluations and debugging failures; familiarity with RL/reward design is a plus; analog circuits/EDA/SPICE/scientific computing is a plus
Experience requirements
At least 3 years of relevant AI / ML engineering experience, with hands-on model training or fine-tuning
Benefits
Equity: 0.10% - 2.00%