Member of Technical Staff, Reinforcement Learning

Building out new LNG plants, HVAC systems, or semiconductor processes rely on hundreds of iterations of feasibility testing (through simulation or real-world lab tests) where different designs (combinations of equipment) and parameters are validated and optimized.

The parameters are often locked in PDFs (datasheets, wiring diagrams, PFDs/P&IDs) and the simulators don't have an easy API to work with. Engineers spend hundreds of hours to annotate and digitize these drawings to run simulations, quote work, build automations, investigate issues in factories, and optimize manufacturing processes. Unfortunately, LLMs and VLMs are still quite bad at analyzing these drawings and using engineering tools (like CAE, TCAD, etc).

Outerport bridges the gap by finding documents from PLMs, extracting structured data from drawings, building a knowledge graph over them, and building autonomous AI agents that can fully automate this R&D process by running simulations and performing design checks.

Outerport is already used at Fortune 500 enterprise companies in manufacturing & industrials to speed up design engineering and simulations.

Our team consists of experts in computer vision, computer graphics, systems software, and AI from companies like NVIDIA and Tulip Interfaces. With backing from Y Combinator and top-tier venture capital firms and angels, we're looking to expand our team to accelerate product development.

We work in SF and Tokyo. and we're a team that enjoy playing sports together and interested in topics ranging from physics research to semiconductor supply chains to beautiful industrial design.

We’re looking for candidates with experience building reinforcement learning-based LLM training pipelines.

As part of our founding team you may:

• Train reinforcement learning-based LLMs to solve tasks in the domain of materials science, chemical engineering, and engineering science
• Integrate simulation tools and real hardware to collect data
• Work with and source data from vendors
• Evaluate, test, and deploy models

Qualifications (none of these are hard requirements)

• Industry experience or projects working on hard problems in machine learning (with Python / PyTorch)
• Being scrappy (getting things done, over theoretical soundness)
• Publication record in venues like ICLR, NeurIPS, ICML, and others
• Good visual taste / an appreciation for aesthetics
• A reputation for having an “engineering mindset”
• Strong communication skills
• Experience working with various engineering simulation tools (e.g. CAE, EDA tools)
• Interest in manufacturing, engineering, EPC, semiconductors, materials science

1. Initial screening (background, past experiences, etc)
2. Technical Challenge (interview) + in some cases, follow-up system design interview
3. Paper Presentation (group interview)
4. (In some cases) Work trial