Kennedy Muriuki

I am a software engineer and AI/LLM specialist with 3+ years of hands-on experience building cloud-native AI solutions, evaluation pipelines, and full-stack systems. I focus on translating complex problems into maintainable, data-driven products that deliver measurable impact.

At Scale AI I optimized prompt design and standardized LLM comparisons across GPT-4, Claude, and Mistral, increasing annotator agreement and automating validation to reduce manual effort. I have built backend tools and APIs using Flask and FastAPI and implemented AI-assisted features that improved processing speed and reduced errors in production settings.

My projects include an LLM Output Scoring Dashboard for real-time evaluation and a CodeGen Assistant VS Code extension with prompt-injection hardening and anonymized telemetry. I consistently apply best practices in testing, automation, and telemetry to improve reliability and performance.

I thrive in cross-functional, remote teams and am committed to continuous learning, leveraging Python, C++, Java, TypeScript, cloud services (AWS, Azure), and modern ML tooling to deliver scalable solutions.