Introduction
Professional Experience
Education
Technical Skills
Certifications
Researcher & Developer
github.com/Anuj-G-06. Code, research repos and open-source work.
linkedin.com/in/anuj-gupta-2k. Experience and recommendations.
@Anuj_M_G. Thoughts on AI, research and building.
Google Scholar profile. Publications and citations.
@anuj.1306.gupta. Longer writing and explainers.
anujg2@cs.cmu.edu. The best way to reach me.
Resume (PDF). A one-page summary of my work.
Researcher and developer with a passion for turning complex challenges into production-grade machine learning solutions.
Experienced across healthcare AI, financial analytics, NLP, large language models, and computer vision, from research prototyping to scalable cloud deployments on GCP and AWS.
Experienced across healthcare AI, financial analytics, NLP, large language models, and computer vision, from research prototyping to scalable cloud deployments on GCP and AWS.
Professional Experience
Aspection LabsSep 2026 - Present
Technical Contributor- Technical contributor at AspectionLabs.com, working on video generation research and applied AI.
Carnegie Mellon University SCSAug 2026 - Present
Graduate Researcher at TEEL Lab- Back at TEEL Lab after my Capital One internship, continuing research on language models for learning and education.
Capital OneJun 2026 - Aug 2026
Senior Data Scientist Intern- Developed a LangGraph agentic system for the Anti-Money Laundering team to automate manual suspicious activity investigations, reducing resolution time by 98% (70 mins/case to 2 mins/case) and directing resources to critical scenarios.
- Engineered features such as customer demographics and prior filings by querying multiple sources including Snowflake, BigQuery, and DBT, improving ETL throughput by 3x with case-based caching systems.
- Benchmarked 300 case variants on the company's internal cloud platform across 3 prompting techniques, 4 multi-agentic architectures, and 2 RAG embeddings, selecting the configuration with the best accuracy-to-false-positive ratio.
Carnegie Mellon University SCSAug 2025 - May 2026
Graduate Researcher at TEEL Lab- Developed LLM benchmarks across finance, healthcare, news, and law domains for semantic annotation of long-form documents, aligning language models with real-world NLP tasks and identifying gaps in existing academic benchmarks.
- The semantic span annotation benchmark from this work was published at ACL 2026.
NeuraLumeJun 2025 - Aug 2026
Founder- Built an AI video generation product end to end, including an agentic pipeline that turns a written story into a finished video.
- Incubated at Carnegie Mellon's Swartz Center for Entrepreneurship, where I built the product's web studio and public site.
EquifaxAug 2023 - Jul 2025
Data Scientist- Analyzed nationwide credit portfolio and feature-engineered account-level parameters for predictive insights, guiding loan disbursement strategies impacting over $45.5B in microfinance portfolios.
- Built automated financial analysis reports using LangGraph and local LLMs to process large-scale market data with multi-step reasoning, improving reporting efficiency by 90% across 300+ risk projects.
- Evaluated Random Forests, Gradient Boosted Models, and XGBoost for interpretable credit loss prediction, deploying XGBoost with SHAP explanations on Vertex AI to achieve 87.6% accuracy.
- Engineered geospatial visualization pipelines in Airflow with GeoPandas and Mapbox APIs to track and forecast regional delinquency trends, driving a 35% improvement in client conversion strategies.
- Led cloud migration of legacy systems to GCP BigQuery, Cloud Dataflow, and Apache Spark, reducing annual operational costs by 40% and improving pipeline runtime by 27%.
- Received 2 Spotlight Awards from global leadership for outstanding contributions to data science projects and innovative ML solutions.
QuantiphiAug 2022 - Jul 2023
Software Developer, Healthcare AI- Developed a Generative Adversarial Network (GAN) pipeline in AWS SageMaker to generate 10,000+ synthetic breast cancer images, driving a 15% increase in early-stage detection coverage in clinical risk models.
- Engineered image feature extraction for lung cancer prediction ensembles using LBP texture analysis and Canny edge detection in OpenCV, upgrading classification accuracy by 45%.
- Optimized SQL-driven data pipelines for an Angular application on AWS EC2 with Auto Scaling and CloudWatch monitoring, resulting in $75,000 annual cost savings.
ExpifyJul 2021 - Jul 2022
Founding Software Engineer- Led a team of 5 engineers building a scalable Next.js platform with MongoDB and Razorpay API payment integration, growing user acquisition by 85%.
- Developed a Word2Vec-based clustering algorithm on psychometric test datasets using Gensim, NLTK, and scikit-learn, achieving 79% accuracy in career prediction for new customers.
Nippon Life India Asset Management LimitedMar 2021 - Jul 2021
AI Research Intern, NLP & FinTech- Fine-tuned a 340M parameter FinBERT transformer with Named Entity Recognition (NER) in PyTorch for automated stock-price-to-sentiment correlation reports, refining investment decision metrics by 1.5%.
- Built real-time web scraping pipelines with Selenium and regex for news data extraction powering NLP sentiment analysis models.
Education
Carnegie Mellon UniversityMaster of Computational Data Science (MCDS), School of Computer ScienceResearch at the TEEL Lab on LLM benchmarking and semantic annotation.
Fall 2026
Now- 10-703Deep Reinforcement Learning and ControlKaterina Fragkiadaki, Max Simchowitz
- 11-632Data Science CapstoneKemal Oflazer
- 11-635Data Science Capstone ResearchKemal Oflazer
- 11-777Multimodal Machine LearningLouis-Philippe Morency
- 11-801Quantitative Evaluations of Language TechnologiesFernando Diaz
My final semester: deep RL, multimodal models, evaluating language technologies, and a capstone on code agents.
Summer 2026
Senior Data Scientist Intern at Capital One.
Spring 2026
- 11-634Data Science Capstone Planning SeminarKemal Oflazer
- 11-868Large Language Model SystemsLei Li
- 15-619Cloud ComputingSeth Goldstein, Majd Sakr
A systems semester. Cloud Computing had me shipping containers and Kubernetes services every week, LLM Systems is where my work on faster GRPO rollouts came from, and I started planning my capstone. kvcache-code-rl-optimize
Fall 2025
- 05-839Interactive Data ScienceAdam Perer
- 10-701Machine Learning (PhD)Hoda Heidari
- 11-631Data Science SeminarMaarten Sap
- 11-637Foundations of Computational Data ScienceDavid Mortensen, Carolyn Rosé
- 11-667Large Language Models: Methods and ApplicationsYubin Kim, Jaromir Savelka
My first semester. I went deep on ML theory and on how large language models actually work. The Interactive Data Science final project became RedViz.
IBM x DJSCEMinors in Artificial Intelligence and Machine LearningSpecialized in deep learning, NLP, computational linguistics, and applied statistics. Developed hands-on skills in pattern detection, anomaly detection, and industrial AI/ML applications through projects and lab work.
Dwarkadas J. Sanghvi College of EngineeringBachelor of Engineering in Electronics and TelecommunicationCore foundations in signal processing, digital communication, computer networks, image processing, and database management. Additional coursework in object-oriented programming (Java), microprocessors, and AI-driven technologies.
Technical Skills
Programming Languages
LLM & NLP Frameworks
Machine Learning & Deep Learning
Data Engineering & Databases
Cloud & MLOps
Certifications
Google Cloud CertifiedAssociate Cloud Engineer
AWS CertifiedSolutions Architect Associate