I build intelligent systems that solve real-world problems — agentic AI workflows, LLM applications, and end-to-end ML pipelines that ship.
|
🤖 Agentic AI Systems Multi-agent workflows, reasoning loops, autonomous task execution ⚙️ AI-Powered Process Automation Replacing repetitive office workflows with robust, human-in-the-loop pipelines 🧠 LLM Applications Natural language → SQL, semantic parsing, inference optimization |
🎯 LLM Fine-Tuning LoRA / QLoRA, domain adaptation, rigorous model evaluation 🖼️ Computer Vision Image classification, brain-activity-based image reconstruction 🛠️ ML Engineering End-to-end pipelines, Docker, CI/CD, FastAPI |
⚙️ AI Automation Engineer — automating real-world business operations end-to-end: turning manual, hours-a-day workflows (scheduling data → generated documents) into tested, self-verifying pipelines that save teams 40–50% of their daily workload
🎓 Research Assistant @ USC Center for the Neuroscience of Embodied Cognition (CeNEC) — connecting cognitive science with AI tools and methods
💬 Lead Developer of ChatDB — a natural-language interface to SQL & NoSQL databases powered by LLMs
| Project | What it does | |
|---|---|---|
| 🔧 | Silicon Validation Log-Analysis Agent | LLM agent that digs through hardware validation logs so engineers don't have to |
| ⚡ | llm-inference-optimizer | High-throughput, low-latency LLM inference platform for LLaMA-3 & Mistral |
| 🕵️ | Agentic Anomaly & Root-Cause Intelligence | Agent-driven platform that detects anomalies and reasons about their root cause |
| 💬 | ChatDB | Talk to SQL & NoSQL databases in plain English via LLMs |
| 🧠 | Brain-Thought-Visualization | Reconstructing visual experience from brain activity |
Languages & ML
Deployment & Infra
Data
I write practical deep-dives on LLMs, agents, and ML engineering → medium.com/@SapnilPatel
⭐ If something here interests you — let's talk.