Shadrack Koech

I am a Software Engineer and Agentic AI Trainer with 5+ years of experience building scalable backend systems, RESTful APIs, and data-driven applications. My work spans the full software development lifecycle — from system architecture and database design to deployment and DevOps automation.

Most recently, I worked at Turing as an RL-Gym Developer and Agentic AI Trainer, where I designed reinforcement learning environments using FastAPI and PostgreSQL, enabling AI agents to perform complex multi-step workflows across platforms like Google Drive, Databricks, Asana, and more. I also served as a Pod Lead, managing and mentoring a team of six AI trainers delivering RLHF, SFT, and agentic training pipelines for clients including Meta and ServiceNow.

On the software engineering side, I have built and maintained production-grade backend services using Python (FastAPI, Django, Flask), with PostgreSQL, MySQL, MongoDB, and Redis as data stores. I'm comfortable working across the stack — with React.js, Next.js, and TypeScript on the frontend — and have hands-on experience with Docker, Kubernetes, CI/CD pipelines, Nginx, and cloud-native infrastructure.

I have led full-stack projects end-to-end, including a Material Management System for palm oil tracking and a Dispatch and Returns System, both of which I architected, built, and deployed independently. I bring strong research thinking, clear technical documentation habits, and the ability to thrive under ambiguous goals.

I hold a Post-Graduate Certification in Software Engineering from ALX-Africa (Holberton School) and a BSc in Industrial Chemistry from JKUAT, giving me a unique combination of analytical rigour and engineering discipline.