Nicholas Koester

Technical Anchor — Full Stack Web Developer

Ford Motor Company
Nov. 2022-Present

Large-Language Model Website

  • Designed and implemented Ford internal large-language model chat website using React and Tailwind.
  • Interfaced with OpenAI GPT, Google Gemini, and Meta Llama models to provide LLM capabilities to Ford employees.
  • Designed and implemented internal API to expose LLM capabilities to Ford applications programatically.
  • Scaled Node.js backend to support over 50,000 users and 5,000 daily conversations.
  • Implemented response streaming to reduce response times and improve user-experience
  • Optimized site performance by migrating frontend to Next.js and using server-side rendering.
  • Enforced code-quality and test coverage standards by reviewing code changes.
  • Collaborated with product owners to define, prioritize, and delegate user stories.
  • Integrated NLP website with classification models to label data automatically and implemented active learning workflow to reduce the amount of manual labeling that was needed before a model could be trained.
  • Developed and deployed model catalog where users could publish models to be reused across common data formats.
  • Rewrote testing standards to increase confidence in code by making tests more closely reflect user behavior.
  • Collaborated with product owners to estimate, prioritize, and delegate user stories.

Natural Language Processing Website

  • Created website using React and Node.js to allow users to index, search, and label company data.
  • Designed and implemented model training workflow to train models to label data automatically.

Connectivity Software Research Engineer

Ford Motor Company
Feb. 2019-Nov. 2022

BlueCruise Driver Assistance

  • Implemented initial version of the Connected BlueCruise feature in Ford vehicles in C++.
  • Integrated and debugged first iteration of software in test vehicle to meet vehicle release timeline.
  • Aggregated road feature data in Java to generate SQL database to represent highway map data for North America
  • Received Research & Engineering Technical Achievement award for implementation.

Research Engineering Intern

Ford Motor Company
Jun. 2018-Sep. 2018

Block-Based Script Editor

  • Created block-based script editor desktop application in JavaScript to enable rapid deployment of scripts to Ford vehicles.
  • Implemented abilitiy to convert blocks into Lua code which includes custom data collection APIs.