Arnav Dhar
AI Engineer & Researcher · Durham, NC
Math + CS at Duke. I research and build applied machine learning for health.
I'm building Anchor, a startup bringing AI and custom software to high-end fashion brands, now a $6K/mo client engagement; developing VaultLab, a multi-agent AI research companion for Claude Code; and working as a software engineering intern at PROTECT3D.
Education
Duke University
Aug 2025 – PresentB.S. Mathematics & Computer Science
HackDuke 1st Place · Duke Applied ML Senior Engineer · Startup Showcase Top 10
Northwood High School
2021 – 2025Cum Laude · Irvine, CA · GPA 3.97 · SAT 1560
National Merit Finalist · President's Volunteer Service Award (Gold) · ML Club President
Skills
Languages
- Python · Java · TypeScript · SQL
AI / ML
- PyTorch
- scikit-learn
- Hugging Face
- RAG & evals
Engineering
- FastAPI
- Next.js / React
- Postgres
- Docker
Shipped
- RAG pipelines
- LLM evals & golden tests
- Agentic workflows (MCP)
- Multi-agent orchestration (LangGraph)
Experience & Projects
VaultLab
- An AI research companion for Claude Code that runs the full literature-to-deliverable research pipeline through coordinated multi-agent workflows.
- Orchestrates specialized roles (data analysts, literature surveyors, adversarial critics) across a structured knowledge base, with every finding tracked through tiered citation verification and reproducibility receipts.
- Moves AI-generated research from raw papers to publication-ready figures, reports, and slide decks without sacrificing provenance or rigor.
- Multi-agent LLMs
- Claude Code
- Research automation
AI Consultant
$6K/moAnchor · Durham, NC
- Building Anchor, a startup that brings AI and custom software to high-end fashion and fine-jewelry brands.
- Shipped generative-design tooling that turns gemstone photos into jewelry designs, plus personalized client recommendations.
- Secured and served a $6K/mo contract with Ruchi New York, a fine-jewelry brand with celebrity clientele; led a team of three.
- High-end fashion
- Generative AI
- Consulting
Software Engineering Intern
PROTECT3D · Durham, NC
- Built a multi-stage cascaded XGBoost pipeline predicting custom orthotic design parameters from 3D foot-scan landmarks, at ~3% average error.
- Replaced manual clinical measurement with automated ML inference; adopted across PROTECT3D's product line.
- XGBoost
- MedTech
AI/ML Research Intern
Mind Computing · San Diego, CA (Remote)
- Improved clinical terminology translations for the Knowledge Platform, increasing AI model performance by 23%.
- Built a speech-to-computable-text system for clinical encounters — runner-up in the national AI Tech Sprint among 152 companies.
- Clinical NLP
- Speech AI
Research
Albert Einstein College of Medicine · AI Researcher · May 2024 – Aug 2025 · Remote
- Built an ML model predicting Mild Cognitive Impairment progression to Alzheimer's at 88% accuracy; now extending it into an LLM-based neurodegeneration-prediction system as my undergraduate thesis.
- Co-developed a personalized breast-cancer treatment model (83% accuracy), published in the National Library of Medicine.
- Wrote the lab's standard operating procedure for cloud VM access, now used by 150+ researchers.
- LLMs
- Predictive Modeling
- Published
Cognitive Synergy Corporation
U.S. PatentAI & ML Research Intern · Aug 2023 – May 2024 · San Diego, CA
RespRisk
PublishedIndependent Study · Nov 2024
Hobbies
Outside of work I'm usually on a tennis court, trying new food, or giving Coco belly rubs.
Elsewhere
Reach me directly at arnav.dhar@duke.edu.