Siddhant Gupta

I'm Siddhant Gupta, an AI engineer who likes building systems that feel real, not theoretical. I work on agentic and generative AI products end-to-end. That means model behavior, evaluation, and orchestration, as well as the unglamorous parts that make things usable, reliable, and safe. I care a lot about observability and trust, because if an AI system can't be debugged, it can't be relied on.

A lot of my work sits around LLM systems, RAG, structured outputs, tool calling, and agent workflows. I've built production pipelines that turn messy inputs like long documents and mixed data sources into clean, typed outputs, and I've spent time fine-tuning models and improving behavior through better data, feedback signals, and evaluation loops. I like pushing for simple interfaces on top of complex systems, so teams can move faster without guessing what's happening under the hood.

I also build in public. I maintain open-source projects, write blogs about what I'm learning, and share practical lessons from building agents, improving retrieval, and making AI workflows more predictable. I enjoy participating in hackathons and AI events because I like meeting builders, testing ideas quickly, and shipping things people can actually use. Those spaces keep me honest, because the feedback is immediate and the bar is real.

Outside of work, I'm into football, cooking, and gaming.