Rakshit Sisodia

I am an AI backend engineer specializing in building and operating scalable LLM inference APIs and distributed microservices across cloud platforms. I design systems for low-latency streaming responses, cost-constrained inference, and high-fanout asynchronous workloads.

My work has driven measurable improvements in reliability and performance — from reducing scoring inconsistency by 31% through rubric-first evaluation to improving quiz throughput 42% and cutting p95 search latency by 78%. I deliver end-to-end solutions including caching, sharding, orchestration, and CI/CD to support millions of daily users and high-concurrency workloads.

I contribute to open source projects addressing production-critical issues and build full-stack projects that combine embedding retrieval, long-running inference, and multi-cloud integrations. I seek roles where I can lead backend reliability, optimize ML inference pipelines, and scale systems for enterprise usage.