Rakshesh Rajeshkumar Shah

Led design and development of a latency-sensitive, scalable backend recommendation engine powering personalized suggestions on various Amazon pages. Built and maintained infrastructure for end-to-end ML workflows, prototyped LLM-based solutions for customer experience, improved infrastructure reliability and observability, collaborated cross-functionally to drive experimentation and product impact, mentored engineers, and contributed to adoption of distributed systems best practices and MLOps automation.

Led design and development of latency-sensitive, scalable, distributed backend recommendation engine powering personalized suggestions across Amazon. Built infrastructure supporting end-to-end ML workflows including training, evaluation, and real-time inference. Explored LLM-based solutions for customer experience and personalization specifically in Amazon Beauty and related verticals. Improved reliability, observability, and testing frameworks. Collaborated cross-functionally for roadmap and experimentation, while mentoring engineers and improving team processes.

Spearheaded core initiatives to improve Facebook’s entity search relevance, architected a code-review-time test automation framework to reduce SEVs and enhance code health, designed and deployed advanced LLM-based evaluation pipelines, optimized computational efficiency with intent-driven query optimization, collaborated across product and research teams to innovate semantic search techniques, contributed to LLM-integrated search workflows including RAG and embedding-based search, mentored engineers on scalable ML system design and LLMOps, and drove Meta’s broader AI strategy aligning research with production-grade implementation.