Jan 2024 – Aug 2025 · 1 yr 7 mos · Seattle, Washington, United States

• Designed and developed a Retrieval-Augmented Generation (RAG) based GenAI system that converts natural language into CloudTrail Lake queries using AWS Bedrock, LangChain, and Anthropic Claude, experiencing 30+ TPS traffic and driving a 30% increase in adoption.
• Built and scaled an LLM evaluation framework leveraging text embeddings and scoring logic to validate 2K+ summaries/day, catching regressions before they impacted users.
• Owned on-call for high-impact distributed systems, resolving 20+ tickets per cycle and reducing operational overhead by 15%.
• Mentored new engineers and contributed to improvements in system reliability, observability, and developer efficiency.

Dec 2021 – Dec 2023 · 2 yrs

• Designed and developed large-scale infrastructure solutions to support training and deployment of in-house LLMs, including productionizing a model based on MPT-30B to serve 50+ requests/sec.
• Built robust ETLM pipelines and orchestrated 1,200+ ML workflows/day using AWS Step Functions, improving data freshness and automation.
• Reduced model refresh time by 20%, boosting ML scientist productivity by ~50% and accelerating time to production for new models.
• Led cost optimization initiatives across ML services, reducing model serving costs by 10% while improving scalability and performance.

Jul 2020 – Dec 2021 · 1 yr 5 mos

• Designed and developed 5+ scalable microservices for seller onboarding, registration, and tax workflows, enabling 10,000+ seller onboardings/day.
• Built asynchronous ETLM pipelines to process ~15,000 seller discounts/day, saving 40+ hours/week of manual effort.
• Led cost and resource optimization across 10+ services, resulting in an 8% reduction in infrastructure spend.
• Collaborated with cross-functional teams to drive efficiency in system migrations and improve platform reliability for internal stakeholders.

Sep 2018 – Jun 2020 · 1 yr 9 mos

Jan 2018 – Jul 2018 · 6 mos