Applied AI, machine learning systems, and computer vision
Rahul Kavi
Staff Research Engineer/Scientist at ServiceNow working on enterprise AI agents, LLM evaluation, and production ML systems.
Profile
I build enterprise AI agents and applied machine learning systems, with experience across generative AI, NLP, deep learning, and computer vision. At ServiceNow, my work focuses on agent planning, orchestration, ReAct-style task execution, permission-aware tool use, and evaluation of agent behavior across task success, robustness, latency, cost, and model compatibility.
Previously at KLA, I worked on machine learning and computer vision for semiconductor defect inspection. I hold a PhD in Computer Science from West Virginia University, where my research focused on activity recognition, multi-view sensor fusion, and image classification.
Current Focus
Enterprise AI Agents
Planning, orchestration, tool selection, task execution, and recovery from failed attempts.
LLM Evaluation
Task success, failure modes, cross-model behavior, latency, and cost tradeoffs.
Production ML
FastAPI services, Docker packaging, REST APIs, and product-facing ML integration.
Computer Vision
Deep learning workflows for image classification, defect inspection, and sensor-based activity recognition.
Experience Highlights
- ServiceNow: Developed AI Agent Studio capabilities for planning, orchestration, and task execution; evaluated behavior across supported LLMs.
- KLA: Built deep learning workflows for silicon-wafer defect classification and worked with engineering teams to integrate ML into products.
- OpenCV: Contributed to the open-source computer vision library used widely across research and production systems.
- West Virginia University: Researched computer vision and activity recognition using cameras, wearable sensors, and edge devices.
Tech Stack
| Area | Technologies |
|---|---|
| Languages | Python, C++, Java, JavaScript, Bash |
| Generative AI and Agents | LangChain, LangGraph, MCP, RAG, LLM guardrails |
| ML and Vision | PyTorch, TensorFlow, scikit-learn, OpenCV |
| Data and Retrieval | NumPy, Pandas, SciPy, FAISS, Pinecone, SQL |
| Services and Tooling | FastAPI, Docker, REST APIs, Git |