Remote AI/ML Engineer responsible for building an agentic platform using Agno, LangChain, and LangGraph on Google Vertex AI with Python microservices and BigQuery; standardized tools, memory, and guardrails for reliable, auditable runs. Implemented a Matchmaking Agent to normalize investor preferences and sponsor profiles, rank candidates, and generate rationale and outreach, boosting investor outreach by 4x. Implemented a Due Diligence Agent with RAG/NLP pipelines (spaCy + Transformers) to extract, cross-check, and summarize key terms from pitch decks and documents, reducing manual review effort by 40%. Implemented an Onboarding Agent with automated KYC/KYB intake (OCR documents, entity checks, rules + LLM verification) to pre-approve/flag applicants, cutting handoffs and speeding up approvals by 10x. (Team-tracked KPI).

Graduate Research Assistant - Machine Learning/Deep Learning at University of Maryland, Baltimore County

May 1, 2025 - May 1, 2025

Implemented LSTM-based time series forecasting models to forecast weather, achieving a 40% increase in prediction accuracy.

Data Science Engineer – Intel laSphere Client Project at LSN Software Services Pvt. Ltd.

July 1, 2023 - July 1, 2023

Architected and maintained scalable ETL pipelines in Apache Spark and Airflow to ingest, clean, and transform 100M+ records/day from multi-source marketing datasets into AWS Redshift, reducing latency by 40%. Developed and deployed Transformer-based NLP models (BERT, RoBERTa) in PyTorch for Sentiment Analysis and SEO keyword classification, achieving 92% accuracy and boosting organic search traffic by 30%. Built predictive models for Email Marketing engagement using scikit-learn and PyTorch—segmenting audiences by click-through propensity and increasing open/click rates by 25%. Engineered a Content Recommendation system combining collaborative filtering with contextualized embeddings, delivering personalized suggestions that lifted user dwell time by 18%. Packaged ML models as Docker containers and orchestrated deployment on AWS ECS/EKS; integrated CI/CD workflows with GitHub Actions for reliable, automated releases.