At Sorsiri, I integrated LiteLLM as a centralized AI gateway across multiple LLM providers, with model routing, API keys, and real-time usage and cost monitoring. I also automated billing with Apache Airflow.
I deployed and secured production infrastructure using Docker, Redis, Nginx, and Linux, with Authentik-based OAuth 2.0, OIDC, and JWT authentication, plus MinIO object storage for zero-downtime releases.
As an AI Software Engineer Intern at Sorsiri, I engineered AI backend workflows and integrated LLM-powered services into REST APIs. I also delivered features including prompt engineering, Crawl4AI web crawling, real-time analytics, and Umami user-behavior tracking.
My projects include AlphaLens, an AI equity research platform with forecasting models and a citation-backed research agent, and an offline RAG voice assistant that reduced query latency by 40%. At Engineering College Ajmer, I also applied NLP to UAV communication logs, improving recovery accuracy by 20%.
