Deep Learning Engineer

Posha is building the future of home cookingautonomous robots that live in your kitchen and cook complete meals without human intervention. Powered by AI-driven computer vision, Posha reliably cooks great food, every time

Recognized as one of TIMEs Best Inventions of 2025, Posha has received rave reviews from The Verge and WIRED for bringing robotics and AI into everyday kitchens.

Were a fast-growing Series Afunded startup backed by Accel Partners and were just getting started.

See it in action: https://youtu.be/Ubd6NkZOEkM?si=SFex-isehKHzvZs_

What can I expect from this role?

Posha is one of the rare robotics startups building production-grade consumer robots, with hundreds of units deployed in real homes.

This is a Sr. Individual Contributor role where youll work on large-scale, real-world culinary vision data captured from robots cooking in homes every day. Your mandate is to convert this data into production-grade perception and control systems for autonomous cooking.

What will my work responsibilities look like if I join Posha?

You will own decisions like:

• Training Models: Build and train models that recognize ingredients, prep styles, volumes, and real-time state changes to drive reliable autonomous cooking.
• Autonomy strategy: Define the right autonomy target for Posha, based on real kitchen variability, and translate it into a clear Operational Design Domain for what we support first and what we deliberately do not.
• Data strategy: Decide what data we need, in what order, and why. Set the mix across on-robot data, human demonstrations, and any external data sources, with a strong bias toward data that actually transfers to deployed performance.
• Failure driven learning loop: Define how we use intervention data and edge cases to improve reliability over time, including what we log, how we replay, and how we prioritize the next fixes.
• Model strategy: Choose when to use large general models versus smaller distilled or task tuned models, optimizing for latency, cost, debug-ability, and reliability, not just benchmark performance.

What makes you a match for us?

• Deep expertise in ViT models
• A minimum of 3 years of experience having shipped deep learning models in production

What The World Thinks of Posha

• Posha is comparable to Tesla FSD
• Posha is a fairly good sign that the future is closer than I thought - The Verge
• Poshas tagline is See dinner cook itself, and it largely lives up to this. - WIRED

Why Posha?

• Few organizations offer challenges similar to ours. Building completely autonomous cooking robots requires thinking from first principles and building from ground up.
• We believe that Posha is a generational company with the opportunity to alter the trajectory of how food is cooked inside homes today. This opportunity is parallel to how autonomous cars have made an impact.
• This role will be at the forefront of this opportunity.