At WebOsmotic Private Limited, I design and deploy production Generative AI applications, including customer-facing assistants and real-time multilingual voice agents. I built RAG pipelines with Voyage AI embeddings and Milvus, and cut median response latency by 40% using Redis caching and task-specific model routing.
At Turing, I produced RLHF and supervised fine-tuning data and evaluated GPT and DeepSeek outputs. Earlier, as an Artificial Intelligence Intern at Ernst and Young Global (LLP), I developed computer-vision systems for port operations; my projects include Custama, Barsiq AI, and Assure AI.

