← home

Work

Professional experience, projects, and things I've shipped.

Projects

Market Lens2025

Multi-API financial intelligence system integrating 6+ data sources with intelligent deduplication and LLM-powered sentiment analysis. Auto-syncs to Google Sheets for real-time portfolio signal tracking.

PythonAnthropic ClaudeGrokAlpha VantageGoogle Sheets
→
Deep Learning From Scratch2022

Trained an LSTM on classical guitar compositions to generate novel musical sequences. Also implemented logistic regression, CNNs, and MLPs from scratch without high-level libraries.

PythonTensorFlowmusic21
→
CryptoSimulator2021

Back-end crypto trading simulator with real-time Coinbase exchange data, secure user authentication, two-factor auth, and API key management.

Node.jsExpressMongoDBCoinbase API
→

Experience

PayPal
Software Engineer · San Jose, CA
Jun 2023 – Present
  • Drove cloud-native migration of compliance case-management platform (OneSCM) from on-prem Oracle to GCP (Cloud Run, Spanner, GKE, Pub/Sub); presented cell-based multi-tenant architecture to SVP/CTO leadership.
  • Engineered fraud refund & currency-conversion system — recovers $160M/year from fraudulent accounts, mitigates $10M in litigation risk.
  • Automated SAR filing across 5 AML rules — eliminates 10K+ manual investigations/year, saves 30K+ investigator hours annually.
  • Built Account Bundling tool: consolidates 1M+ transaction records, automates multi-jurisdiction filing.
  • Engineered 3 LLM chat assistants for case processing, team routing, and compliance narrative generation.
  • Built 2 Agentic AI solutions (Crew AI): voice-activated case assistant + Oracle SQL prompt-to-PR generator.
  • Re-architected Law Enforcement Fulfillment Portal (25K+ subpoenas/year): −55% latency and storage, saves $5M/yr.
  • Orchestrated NCMEC integration: automates 2,000+ child-exploitation case reports/year to law enforcement.
JavaSpring BootGCPAWSLangChainCrew AIOracle SQL
PayPal
Software Engineer Intern · San Jose, CA
Jun 2022 – Aug 2022
  • Built Java/Spring ETL connector: extracts and loads 3.5M case-management logs into Celonis for process mining, enabling discovery and operational monitoring.
  • Automated daily movement of thousands of new case records via scheduled batch jobs.
JavaSpring BootCelonisETL
TestAIng
Machine Learning Intern · Bangalore, India (Remote)
May 2021 – Aug 2021
  • Developed ML pipelines for time-series forecasting with concept-drift detection — 7% error reduction on enterprise datasets.
  • Integrated LIME and SHAP explainability methods for non-technical stakeholders.
Pythonscikit-learnLIMESHAP
Arevea Inc.
Software Engineer Intern · Remote
May 2021 – Aug 2021
  • Designed admin portal for a virtual event management platform.
  • Implemented secure RESTful APIs with QR-code-based event check-in functionality.
AngularNode.jsSQL
UMass CICS
Undergraduate Research Assistant · Amherst, MA
Dec 2021 – May 2023
  • Evaluated causal inference algorithms on non-linear observational datasets in R under PhD researcher Pracheta Amaranath.
  • Investigated LIME/SHAP explainability of NLP search engines (LETOR) under Prof. James Allan at CIIR.
RPythonLIMESHAPNLP