Experience & Skills
Industry, research, and teaching roles
Industry
Senior Data Scientist — Generative AI
Sep 2025 – Jan 2026
Part-time CPT Internship New York City
- Developed a reasoning-driven Python financial agent for automated credit analysis using LangGraph-based orchestration
- Reduced manual analyst computation and validation time by 60% through optimized agent execution and tool chaining
- Enhanced the SQL agent, improving execution accuracy by ~15% on Scalata's financial benchmark
- Fine-tuned Llama 3.1 8B for financial credit analysis using distributed training on Amazon SageMaker
- Designed a human-in-the-loop pipeline for intent classification and agent orchestration, routing financial queries to the right LLM workflows and improving intent classification accuracy by ~10%
- Implemented multi-agent coordination with LangChain, LangGraph, and LangSmith for production-grade financial AI
- Designed a Knowledge Graph for Scalata Personalized Recommendation using AWS NeptuneDB
Senior Data Scientist — Generative AI
Jun 2025 – Aug 2025
Full-time CPT Internship New York City
- Designed an intelligent data lake supporting 24+ document types with fast vs. slow ingestion for scalable multimodal processing
- Integrated agentic reasoning for data cleaning and validation — 21% cleaner downstream datasets
- Built an end-to-end financial agent as the backbone of the Exchange Robotics Orchestration System
- Developed VectorDB-backed RAG pipelines for retrieval over heterogeneous financial documents
Research
Research Assistant
Sep 2022 – Present
Big Data Analytics Lab NSF & ONR Funded
- Designed human–LLM agent orchestration frameworks for data science pipelines with customizable quality–cost trade-offs — 0.91 accuracy vs. 0.86 for Auto-Sklearn on model refinement
- Developed ASMR, an agentic AI framework for structured schema generation from unstructured ship-maintenance reports — nearly doubling schema informativeness
- Designed a multi-agent LLM framework for personalized recommendation — 45% speedup, 15–20× cost reduction
- Developed probabilistic package-selection algorithms over multimodal data that minimize expensive LLM calls while guaranteeing optimal or near-optimal solutions
- Developed APEX, an algorithmic framework for progressively evaluating agentic workflows to identify potential failures
- Proposed G-CLAHE: 17% improvement in diagnostic accuracy for chest X-ray abnormality detection
Teaching
BNFO135 — Programming for Bioinformatics
Teaching AssistantJan 2024 – May 2024
BNFO135 — Programming for Bioinformatics
Teaching AssistantSep 2023 – Jan 2024
CS113 — Programming with Java and MATLAB
InstructorJan 2023 – May 2023
CS101 — Introduction to Computing
Teaching AssistantSep 2022 – Nov 2022
Computational Intelligence
Teaching AssistantSep 2021 – Jan 2022
Operating Systems
Teaching AssistantSep 2021 – Jan 2022
Artificial Intelligence
Teaching AssistantSep 2021 – Jan 2022
Operating Systems
Teaching AssistantFeb 2021 – Jun 2021
Algorithm Design and Analysis
Teaching AssistantFeb 2021 – Jun 2021
Theory of Languages and Automata
Teaching AssistantSep 2020 – Jun 2021
Data Structures
Teaching AssistantSep 2020 – Jan 2021
Discrete Mathematics
Teaching AssistantFeb 2020 – Jun 2020
Advanced Programming
MentorFeb 2020 – Jun 2020
Discrete Mathematics
Teaching AssistantSep 2019 – Jan 2020
Technical Skills
Programming
LLMs & Agents
ML & Deep Learning
Databases
Cloud & Data
Tools