ML Research Scientist, EndoRISE Program

• Applying ML expertise to women's health research using de-identified endometriosis patient data in collaboration with the Courtois Single Cell Biology Lab, focusing on patient stratification and correlation between surgical inflammation markers and immune comorbidities using clustering and association analysis on EPHect-standardized clinical data

Principal Machine Learning Engineer

May 2022 – Apr 2025

• Led 12-person cross-functional team through 0-to-1 architecture, development, and production launch of a personalized recommendation system for Autodesk App Store; drove a 13% lift in app downloads across ~708K DAU at <100ms latency
  • Owned end-to-end system architecture for real-time ML inference at scale using Wide & Deep Networks with online feature serving; achieved <50ms p95 inference latency
  • Built feature store and data migration pipeline; reduced model training time by 40% and infrastructure costs by 25%
  • Mentored an intern in building functional prototype; guided algorithm selection and system design decisions
• Led 0-to-1 design and implementation of ML-powered customer support insights platform that proactively surfaced product issues and enabled natural-language querying across ~10,000 quarterly tickets
  • Piloted BERT-based topic modeling framework in partnership with Autodesk Platform Services; surfaced 7 critical product issues whose targeted fixes reduced quarterly support volume by 10.7%
  • Extended system with a RAG architecture layered on vector embeddings, FAISS, and cross-encoder re-ranking for contextually relevant insights beyond predefined topics
  • Redesigned from a single-team POC into a modular platform with configurable data connectors; enabled 3 additional product teams to adopt automated ticket analysis with minimal engineering overhead
• Served as organizational Privacy Champion; led privacy impact assessments across ML initiatives, enforced data retention and ethics standards, and partnered with legal on open-source compliance

Senior Machine Learning Engineer

Jan 2019 – Apr 2022

• Engineered high-performance data compression framework that reduced API logs (~1.4GB/hr) by 99.8%, enabling scalable pipelines for:
  • Anomaly detection to reduce API usage abuse and obtain granular customer behavior insights
  • Driving strategic decisions on API monetization, product subscriptions, and customer retention
• Contributed machine learning expertise to two pro bono projects for Autodesk Foundation; delivering predictive modeling and analytics for Hope Street Group and Yuno Technologies
• Championed Data-as-a-Service (DaaS) adoption across engineering teams; raised Blameless Postmortem completion from 20% to 75% and established self-service analytics for incident learning

Data Science Intern

May – Aug 2018

• Developed XGBoost models to forecast sprint velocity and burn rates for proactive agile coaching interventions (12.3% organization-wide burn rate improvement)
• Built an interactive GitHub analysis tool for real-time tracking of employee contributions to Autodesk's open-source projects, fostering developer community engagement

Graduate Teaching Assistant

Aug – Dec 2018

• TA for CS 4476: Introduction to Computer Vision (165 students) under Dr. Devi Parikh
• Coordinated with the teaching team, graded assignments and projects in MATLAB/Python, and held office hours

Software Intern

Aug 2015 – Feb 2017

• Developed web applications using agile methodology, focusing on client-side programming (HTML, CSS, JavaScript, PHP)
• Resolved and tested 100+ user issues during post-deployment maintenance support across 4 projects over 2 years

Project Trainee

Jun – Jul 2015

• Designed a secure browser-based portal (Epic Essentials) for the Strasz Exam Management Suite, enabling certification and testing organizations to deliver better tests more efficiently

Graduate Student Researcher

Advisor: Dr. Dobromir Rahnev

• Facilitated interdisciplinary AI/neuroscience comparative study research by designing hybrid image classification tasks for models and human responders
• Optimized deep CNN architectures (VGG-16/19, ResNet50, InceptionResNetV2) for hybrid image classification, achieving 94% accuracy through systematic hyperparameter tuning

Undergraduate Researcher

Advisor: Prof. Varunakshi Bhojane

• Publication: Floor Layout Planning using Artificial Intelligence Technique · IJIRSET, Vol. 6, Issue 4, April 2017
• Prototyped Genetic Algorithm software for furniture placement optimization, achieving 90% accuracy in space utilization by processing 20,000 positional chromosome configurations across 5 generations