Luke Bechtel
Founding Engineer at On Deck
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolLuke Bechtel is a founder and AI/ML engineer with 12 years of experience building developer-facing products, realtime 3D systems, and agentic data tooling from San Francisco. He has led and scaled engineering teams through productization and acquisition (Collider Tech → Essentium), driven SOC2-secure cloud architectures, and recently shipped tools for agent workflows and hybrid RAG in data engineering contexts. A hands-on contributor to open-source projects like trimesh and LangchainJS, he focuses on robust backend features—mesh processing, token accounting, and structured output parsing—that improve developer ergonomics. Luke blends product instincts with deep technical chops across computational geometry, ML, and distributed systems, and brings a playful, improv-influenced approach to problem solving that favors experimentation and positive-sum outcomes.
12 years of coding experience
9 years of employment as a software developer
(Inc.) Master's Degree in Computer Science, Machine Learning Specialization, Computer Science, (Inc.) Master's Degree in Computer Science, Machine Learning Specialization, Computer Science at Georgia Institute of TechnologyComputer Science, Computer Science at University of Tennessee, Knoxville
English
Github Skills (17)
mesh10
javascript10
geometry10
python10
typescript10
scikit-image10
numpy10
typescript-types10
typescripts10
zod9
networkx9
openai9
triangulation8
triangular8
triangle8
Programming languages (17)
Github contributions (5)
mikedh/trimesh
Feb 2016 - Oct 2017
Python library for loading and using triangular meshes.
Role in this project:
Backend Developer / Library Contributor
Backend Developer / Library ContributorContributions:7 commits, 4 PRs, 46 comments in 1 year 8 months
Contributions summary:Luke primarily contributed to the `trimesh` Python library by adding new functionalities and improving existing ones. Their work included adding a vertex adjacency graph, a marching cubes mesh generator to the voxel module, and enabling the handling of `Trimesh` subclasses. They also addressed build dependencies by adding `scikit-image` to the required packages, and fixed a typo. The user also added tests to validate the new features implemented.
pythonmeshtriangular-meshesgeometry
langchain-ai/langchainjs
Feb 2023 - Mar 2023
The agent engineering platform
Role in this project:Back-end Developer
Back-end DeveloperContributions:1 review, 3 commits, 10 PRs in 6 days
Contributions summary:Luke contributed to the LangchainJS project by implementing new features related to token counting for chat models and improving the structured output parser. Their work included adding a `getNumTokensFromMessages` function, integrating it with existing model structures, and fixing formatting and test issues related to the structured output parser. These changes suggest the user focused on improving the core functionality of the library, particularly around the chat models and output parsing, leading to a more robust and reliable LangchainJS implementation.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.