Sameer Mohan

I’m a results-driven LLM Engineer with 4 years of experience designing, deploying, and optimizing LLMs and NLP solutions in production environments. My focus is turning fast-moving model advancements into reliable systems that perform under real constraints.

At Together AI, I designed and implemented scalable LLM integration frameworks inside Dataiku, including model routing, prompt orchestration, retrieval-augmented generation (RAG), and agentic AI workflows. I also optimized LLM serving and evaluation frameworks, improving performance, reliability, and cost-efficiency through multi-provider integrations (OpenAI, Anthropic, HuggingFace, etc.).

Previously at Cohere, I joined as an early technical team member and helped build core LLM infrastructure. I designed and scaled training and inference pipelines for large-scale distributed GPU clusters, and I shipped production-grade capabilities like embeddings, rerankers, multilingual systems, and RAG for the company’s API platform.

Through projects, I continue to sharpen my engineering and governance mindset. I’ve built an enterprise multilingual LLM development approach supporting 50+ languages and worked on a Governed GenAI Platform with secure RAG and agentic workflows using governance and guardrails to reduce hallucination rates and accelerate adoption.