Agentic AI Systems
Planning, tool-use, and multi-step workflows that move from chatbot demos to production operators.
Principal AI Applied Scientist
Building production-grade LLM and RAG systems that scale.
I lead applied AI initiatives at Dell, shipping enterprise copilots across search, retrieval, and evaluation pipelines for global support workflows.
- $600K annual LLM cost savings via model migration + compression
- 100K+ internal docs indexed for enterprise search quality
- 2 KDD 2025 publications on retrieval and multi-agent pipelines
About
AI scientist focused on measurable business impact.
I specialize in enterprise LLM systems, retrieval engineering, model fine-tuning, and AI evaluation. My work combines research rigor with shipping mindset: faster support resolution, stronger answer quality, and lower serving cost.
At Dell Technologies, I currently work as Principal / Lead AI Applied Scientist in the Corporate Strategy Office, leading RAG architecture, LoRA fine-tuning, LLM compression experiments, and quality instrumentation with Langfuse and RAGAS.
Areas Of Interest
What I am building now.
Retrieval-Augmented Generation
Hybrid retrieval, chunking strategies, re-ranking, and grounding methods for enterprise knowledge bases.
LLM Efficiency
LoRA fine-tuning, quantization, and compression to deliver lower latency and materially lower inference spend.
Evaluation + Observability
LLM quality gates with RAGAS, Langfuse traces, and experiment loops that connect quality to user outcomes.
Skills
Production stack across modeling, serving, and cloud.
LLM + RAG
ML + Deep Learning
Engineering + Infra
Experience
Recent work highlights.
2024 - Present
Dell Technologies, Corporate Strategy Office
Principal / Lead AI Applied Scientist
- Led enterprise RAG and copilot architecture serving 8,000+ technical support agents.
- Fine-tuned open models using LoRA and built compression workflows to optimize latency and cost.
- Instrumented LLM quality with Langfuse + RAGAS and ran iterative retrieval/evaluation experiments.
- Drove migration from closed APIs to open-weight stack, reducing annual spend by approximately $600K.
2023 - 2024
Dell Technologies
Graduate Intern, Data Science
- Analyzed 1B+ job postings with distributed workflows to detect technology adoption shifts.
- Built lead-targeting frameworks for cloud-to-on-prem transitions and identified priority accounts.
2022
Meesho
Product Analytics Manager, Data Science
- Built ad bidding optimization loops and personalization models, improving monetization outcomes.
Publications
KDD and applied research contributions.
Hybrid Retrieval for Enterprise-Scale RAG Systems
KDD 2025
Multi-Agent BERT for Coordinated LLM Reasoning
KDD 2025
MODP: Multi-Objective Directional Prompting
Patents
Applied IP in LLM retrieval pipelines.
Dynamic Adapter Selection
Patent filing for adaptive routing across LLM adapters in enterprise task orchestration.
Intelligent Chunking Framework
Patent filing for context-aware document chunking to improve retrieval quality and grounding fidelity.
Awards
Recognition across industry and community programs.
TSIA Star Award
TAG Master Modellers
ATLytics Data for Hope
Hacklytics 2023
Projects
Featured work across LLM systems and applied ML.
MODP: Multi-Objective Directional Prompting
Prompt optimization framework for production support systems, delivering measurable quality lift.
Contact