Michael Foster
“Senior Software Engineer | AI | Full Stack | DevOps”
- Joined
- Jun 2023
- How they joined
- Directly
- Before Cognition
- Sensity AITechnical Tead
- Now
- Fireworks AI
- Founding Software Engineer at Cognition, Jun 2023 - Jul 2024. His own role description: "Architected and optimized web data extraction and distributed scraping platforms processing 1M+ pages daily, built LLM-assisted workflows to reduce manual maintenance, implemented QA tools to improve data quality, and collaborated with founders to launch automation features."Jun 2023
- Cognition's acquisition of Windsurf (the agentic IDE, later Devin Desktop lineage) was announced 2025-07-14, about a year after his LinkedIn shows him leaving Cognition (Jul 2024); his public record therefore does not support him working on Windsurf or the acquisition itself.Jul 14, 2025
- His Cognition stint spanned Devin's public debut: Cognition introduced Devin as "the first AI software engineer" in March 2024, roughly nine months into his tenure.Mar 2024
- Cognition was founded by Scott Wu, Steven Hao and Walden Yan, all IOI gold-medalist competitive programmers; as a founding engineer he 'collaborated with founders', meaning these three were his co-founder counterparts at the company.2023
- Incongruity: his LinkedIn dates the Cognition role from Jun 2023, while Wikipedia gives Cognition's founding as August 2023 (other sources say November 2023).Jun 2023
- Sensity AI (deepfake/visual-threat detection): Senior AI Full Stack Engineer Sep 2021 - Sep 2022, then Technical Lead Sep 2022 - Jun 2023, building LLM-powered automation workflows and leading React/TypeScript/Python/Node platforms; he went to Cognition straight from this lead role.Sep 2021
- Lead Software Engineer at EliseAI Feb 2018 - Sep 2021, leading architecture of AI-powered, HIPAA-compliant healthcare communication platforms (microservices, conversational AI, Python/Node/PostgreSQL/Redis) and mentoring engineers.Feb 2018
- Software Engineer at Microsoft Jun 2016 - Feb 2018, building cloud enterprise platforms with C#, .NET, Azure and React plus CI/CD pipelines; no specific Microsoft team or org is named publicly.Jun 2016
- Full Stack Developer at Bitovi (JavaScript consultancy) May 2012 - May 2016, building reusable component systems and REST integrations for enterprise clients; this is the start of the career he describes as "14+ years".May 2012
- Education: Bachelor's degree in Computer Science, Florida Agricultural and Mechanical University (FAMU), 2008-2012.2012
- LinkedIn post (2026-09-08), verbatim opening: "After 10+ years in software engineering, I've learned some important technical principles: - \"Temporary code\" is usually the longest-serving member of the team." It closes "Junior me: \"I can build it!\" Senior me: \"Can we roll it back?\"", showing an operability-first worldview.Sep 8, 2026After 10+ years in software engineering, I’ve learned some important technical principles: • “Temporary code” is usually the longest-serving member of the team. • Every “quick fix” comes with three meetings, five new tickets, and a surprise side quest. • “Works on my machine” is not a solution. It’s a confession. 😂 • The bug is always on the one line everyone agreed was “too simple to fail.” • If
- LinkedIn post (2026-08-15), verbatim: "The best senior engineers don't just solve problems—they prevent them." and "Senior engineering is not measured only by the amount of code written." Neither of his visible recent posts mentions Cognition, Devin or Windsurf.Aug 15, 2026The best senior engineers don’t just solve problems—they prevent them. Writing code is only one part of software engineering. The deeper responsibility is thinking ahead. Before a feature reaches production, a senior engineer asks: Will it scale when traffic grows? What happens when a dependency fails? Is it secure? Can the team maintain it a year from now? Are we solving the real business problem
Built scalable AI data pipelines, developed full-stack features with React and FastAPI, optimized backend performance and AWS infrastructure, led cross-functional teams, and implemented secure multi-tenant systems with authentication and audit logging
Architected and optimized web data extraction and distributed scraping platforms processing 1M+ pages daily, built LLM-assisted workflows to reduce manual maintenance, implemented QA tools to improve data quality, and collaborated with founders to launch automation features.
Led development of AI-driven enterprise platforms with React, TypeScript, Python, and Node.js, architected scalable backend and cloud infrastructure
Built LLM-powered automation workflows, and guided engineering teams through planning, mentorship, and code quality improvements.
Led the architecture and development of AI-powered, HIPAA-compliant healthcare communication platforms — building scalable microservices, conversational AI workflows with LLM integration, and secure real-time infrastructure using Python, Node.js, PostgreSQL, and Redis while mentoring engineers on distributed systems and AI best practices.
Built and optimized scalable cloud-based enterprise platforms at using C#, .NET, Azure, and React — delivering high-performance APIs, reusable full-stack components, and streamlined CI/CD pipelines while contributing to architecture, mentorship, and cross-functional collaboration.
Delivered scalable, high-performance web applications at Bitovi by building reusable component systems, integrating REST APIs, and optimizing legacy codebases to drive 30%+ efficiency gains across enterprise and customer-facing projects.
After 10+ years in software engineering, I’ve learned some important technical principles: • “Temporary code” is usually the longest-serving member of the team. • Every “quick fix” comes with three meetings, five new tickets, and a surprise side quest. • “Works on my machine” is not a solution. It’s a confession. 😂 • The bug is always on the one line everyone agreed was “too simple to fail.” • If
The best senior engineers don’t just solve problems—they prevent them. Writing code is only one part of software engineering. The deeper responsibility is thinking ahead. Before a feature reaches production, a senior engineer asks: Will it scale when traffic grows? What happens when a dependency fails? Is it secure? Can the team maintain it a year from now? Are we solving the real business problem
