Hi, I’m Kunj Rathod, an AI engineer and researcher working on LLM/RAG systems in healthcare and legal tech.

No. KR-2026.06Portfolio · Preprint · June 2026

Systems that retrieve, reason, & remember.

Kunj Rathod1,2

1 Kahlert School of Computing, University of Utah · 2 Microsoft Fabric, Azure Data (SWE Intern, Jan 2026–)

Abstract

AI engineer and researcher working where production systems meet open research problems. Shipped a HIPAA-compliant LLM platform serving 90+ hospital executives on AWS Bedrock; scaled hybrid legal retrieval to 10M+ documents and 5,000+ daily queries; built graph-augmented pipelines for aerospace materials discovery with NASA and DoD collaborators. Current research targets memory for embodied agents — Video Mind Palace cuts long-horizon QA inference time 31–57% by replacing scene-graph world models with direct video-level VLM queries. Now building distributed data systems for Microsoft Fabric, Azure’s Spark-native lakehouse analytics platform.

Keywords: retrieval-augmented generation · embodied agents · long-horizon memory · distributed systems · interpretability

Fig. 0 — Episodic memory lattice, the world model behind Video Mind Palace. Drag to rotate.

10M+

legal docs indexed

<200ms

p95 TTFT, prod LLM chat

31–57%

embodied-QA inference cut

1M+

biomedical graph entities

§1

Experience

industry, research, campus — 2024 to present

Jan 2026 —

Redmond, WA

● active

Software Engineer Intern — Microsoft Fabric (Azure Data)

Microsoft

  • Building distributed data systems for Microsoft Fabric — Microsoft's unified, Spark-native lakehouse analytics platform (the Databricks class of system): OneLake storage, data engineering, warehousing, and real-time intelligence under one compute fabric.
  • Full-stack work across the Fabric platform within the Azure Data organization.

[Microsoft Fabric] [OneLake] [Spark] [Distributed Systems] [Azure]

Jan 2025 —

Salt Lake City, UT

● active

Software Development Intern, AI Services (SUDO Program)

University of Utah Health

  • Built and deployed a HIPAA-compliant AI chat platform for 90+ hospital executives using React/TypeScript, Flask middleware, and AWS Bedrock microservices with event-driven Lambda orchestration.
  • Shipped 6 full-stack features across 4 sprints; integrated AWS Bedrock Agents, Knowledge Bases, and Guardrails for production clinical workflows.
  • Reduced inference latency by 40% and data query speed by 60% via Bedrock pipeline optimization, API caching, and a DynamoDB–RDS hybrid database strategy.
  • Implemented token-streaming LLM responses (p95 <200ms TTFT) with resilient fallback handling and distributed session persistence for 1,000+ conversations.
  • Integrated interactive data visualization tools into the LLM chat interface enabling real-time analytics on hospital data.

[React] [TypeScript] [AWS Bedrock] [Lambda] [DynamoDB] [Flask] [HIPAA]

Jan – May 2026

Salt Lake City, UT

Research Project — Long-term Active Embodied QA

Advanced AI, University of Utah

  • Developed Video Mind Palace (VMP), an efficient agent for Long-term Active Embodied Question Answering (LA-EQA) that replaces scene-graphs with direct video-level VLM queries.
  • Demonstrated a 31–57% reduction in online inference time per query with minimal accuracy trade-offs compared to state-of-the-art Robotic Mind Palace (RMP).
  • Conducted a comprehensive analysis of the LA-EQA benchmark, identifying key limitations in interactivity and proposing future directions using SceneSmith-generated environments.

[Qwen3-VL] [Vision-Language Models] [MuJoCo] [SceneSmith] [Robotics]

Aug 2025 – Feb 2026

Salt Lake City, UT

Undergraduate Researcher — LLMs & Computational Simulations

STARS Lab, University of Utah · with NASA, Microsoft, U.S. DoD

  • Built a multi-agent, graph-augmented pipeline to extract and normalize material-property data from 1,000+ materials-science papers into a physics-aware graph for automated Ashby plot generation.
  • Developed a constraint-based 'design region' engine (temperature, creep, pressure limits) and benchmarking suite to identify feasible materials for extreme aerospace environments.
  • Built Ref-RAG, a custom RAG chatbot using LangChain and Chainlit to extract structured information from large unorganized PDF datasets for materials researchers.

[Python] [LLMs] [Multi-Agent] [LangChain] [Graph RAG]

Nov 2024 – Apr 2025

Remote

AI Engineering Intern

CourtEasy.ai / Nugen

  • Scaled hybrid legal-document retrieval to 10M+ indexed Indian legal documents (statutes, court orders), supporting 5,000+ daily queries.
  • Improved retrieval accuracy by 28% and reduced hallucinations by 35% via hybrid RAG (dense vectors + BM25 + reranking) and context-grounding optimizations for Legal-NER tasks.
  • Built production ETL ingesting 500k+ documents/week and benchmarked 8 LLM families on 4 legal benchmarks including LegalBench and NyayaAnumana; analysis guided model routing, reducing projected inference spend by $50k+/year.

[RAG] [BM25] [Legal-NER] [LegalBench] [Python] [LLMs]

May – Aug 2024

Salt Lake City, UT

AI Research Intern — BioGraphRAG

Garje Marathi Global (GMG Summer of Code)

  • Led development of BioGraphRAG: a Graph Retrieval-Augmented Generation platform combining biomedical knowledge graphs with LLMs for explainable biomedical Q&A.
  • Engineered distributed GraphRAG system managing 1M+ biomedical entities (proteins, genes, diseases) integrating UniProt, AlphaFold, and RXNav with NebulaGraph.
  • Improved factual accuracy by 40%; optimized graph traversal 3× through strategic caching and high-degree node pruning, achieving sub-500ms query latency at p95.
  • Presented at an international AI panel attended by experts from India and the US — received commendation for technical leadership.

[Python] [NebulaGraph] [LlamaIndex] [GraphRAG] [Docker] [FastAPI]

Jan – Apr 2025

Salt Lake City, UT

Campus Strategist

Perplexity AI

  • Spearheaded campus-wide outreach driving adoption of Perplexity's AI search among students, faculty, and clubs; onboarded 150+ Perplexity Pro users.

[AI Advocacy] [Community] [Growth]

Aug – Dec 2024

Salt Lake City, UT

Community Advisor

University of Utah Housing & Residential Education

  • Provided conflict mediation, crisis response, and student support services for a 200+ resident community.

[Leadership] [Crisis Management]

1.1 Honors

Mar 2026

Kahlert Impact Prize $1,000

Kahlert School of Computing, University of Utah

Awarded for societal impact through AI research and production systems in healthcare, legal-tech, and embodied AI. Funded by a $15M endowment from The Kahlert Foundation; recognizes students translating computing research into real-world benefit.

1.2 Education

Aug 2023 – Dec 2026

B.S. Computer Science

University of Utah · GPA 3.7/4.0, Dean’s List

Machine Learning · Computer Vision · NLP · Distributed Systems · Algorithms & Data Structures

§2

Selected Work

figures 01–10, in order of recency
Fig. 01 — interactive
Research

Video Mind Palace

Efficient long-term active embodied QA

  • Agent for LA-EQA that replaces expensive scene-graph world models with direct video-level VLM queries.
  • 31–57% reduction in online inference time per query, within 13–17% of state-of-the-art RMP accuracy.
  • Eliminates mandatory offline GPT-4o captioning preprocessing, cutting end-to-end latency and cost.

Qwen3-VL · MuJoCo · SceneSmith · Python · Robotics

env = SceneSmith("apartment-0042")
rollout = palace.act(env, policy=π)

Live plate — CCD inverse kinematics at 60 Hz. The arm tracks its target autonomously; drag on the floor plane to take over the policy.

Fig. 02
Local AI System

FNDR

Privacy-first local AI memory for macOS

  • Zero-trust, local-only memory assistant in Rust/Tauri — no cloud, no telemetry, full data sovereignty.
  • Metal-accelerated on-device inference (Llama 3.2, SmolVLM) for low-latency RAG on Apple Silicon.
  • Real-time screen extraction via Apple Vision OCR + CLIP embeddings; Graphiti-style temporal search across activities, web sessions, and meeting transcripts.
  • Local Whisper transcription and an MCP server for secure interop with external agents and IDEs.

Rust · Tauri · Metal · Llama 3.2 · Whisper · MCP

Fig. 03
AI Platform

HirePilot

Autonomous AI recruiting agency

  • Fully autonomous recruiting backend with specialized agents — Enrichment, Scheduling, Interview, Evaluation — managing the hiring lifecycle from GitHub sourcing to live screening.
  • Twilio real-time voice AI interviews, Google Calendar slot scheduling, Slack/Resend approval and outreach flows.

TypeScript · Node.js · PostgreSQL · Anthropic API · Twilio

Fig. 04
Application

CloudCoder

Prompt → deployed AWS application

  • Model-agnostic orchestrator embedded in this site: generates and deploys serverless AWS apps directly to a live AWS account.
  • Emits structured React SPAs, Node.js Lambdas, and SAM CloudFormation templates via the Vercel AI SDK.
  • Packages Lambda binaries with JSZip, stages S3 artifacts, and executes CloudFormation with SSE log streaming to the UI.

Next.js · Claude / GPT-4o · AWS SDK v3 · CloudFormation

Fig. 05
Hackathon Winner

Minute0

AI-powered deployment monitor

  • Tracks Vercel deployments, classifies build/runtime failures, and triggers Slack alerts with approval workflows.
  • AI root-cause analysis over logs with FastAPI + ChromaDB vector search, emitting structured fix suggestions for coding agents.

React · FastAPI · ChromaDB · Cerebras · Slack API

Fig. 06
AI Orchestrator

Omni

Everything, everywhere, all at once

  • Unified intelligence layer over Gmail, Google Calendar, Slack, and FNDR private memory — Smart Todos, natural-language scheduling, on-demand personal context.
  • Real-time voice interaction and autonomous multi-step workflow orchestration across the digital stack.

React · TypeScript · MCP · Gmail API · Slack API

Fig. 07
Research System

BioGraphRAG

Biomedical knowledge-graph retrieval

  • Distributed GraphRAG unifying UniProt, AlphaFold, RXNav, and BioKG in NebulaGraph; ETL processes 2M+ entity updates monthly.
  • +40% factual accuracy; 3× faster graph traversal via caching and high-degree node pruning (sub-500ms p95).

Python · NebulaGraph · LlamaIndex · FastAPI · Docker

Fig. 08
Application

Wingman.ai

Multi-modal personal assistant for iOS

  • Voice, chat, and image input over GPT-4o and Whisper with RAG-enhanced memory.
  • Offline-first architecture with Firebase sync, real-time streaming, persistent history.

SwiftUI · GPT-4o · Whisper · Firebase

Fig. 09
System

FlowVía

V2X urban mobility optimization

  • V2V/V2I/V2N platform for real-time adaptive traffic management, from OBD-II hardware to cloud ML backend.
  • LSTM traffic-flow prediction on live SPaT signal data; AES-256 comms with rotating vehicle identifiers.

Python · TensorFlow · LSTM · DSRC · C-V2X

Fig. 10
HackUSU 2025

RL Investment Advisor

Reinforcement-learning portfolio optimizer

  • DistillBERT sentiment analysis on financial news combined with DQN and PPO for portfolio optimization; measurable outperformance on backtests.

Python · DistillBERT · DQN · PPO

§3

Writing & References

  1. [1]

    K. Rathod, N. Subedi. Exploring Long-term Active Embodied Question Answering in Simulated Indoor Environments.” Advanced AI, University of Utah, Spring 2026.

    Investigates the LA-EQA benchmark and proposes Video Mind Palace (VMP), an efficient agent replacing structured scene-graphs with direct VLM queries — 31–57% reduction in inference time with minimal accuracy drop, plus a structural critique of benchmark interactivity and episode length.

  2. [2]

    K. Rathod, N. K. Singh. BioGraphRAG — Biomedical Knowledge Graph Retrieval Augmented Generation.” Kunj's Substack · GMG Summer of Code, Oct 2024.

    System architecture, the GraphRAG algorithm, node-degree performance analysis, and a multi-stage answer-enrichment pipeline integrating UniProt, AlphaFold, and RXNav.

    NebulaGraph's marketing team requested republication on their official website (Jun 2025).

  3. [3]

    K. Rathod. FlowVía: A Technical Deep Dive into Next-Gen Urban Mobility.” Kunj's Substack, Apr 2024.

    V2V/V2I/V2N protocols, DSRC and C-V2X standards, real-time speed recommendation algorithms, LSTM traffic-flow prediction, and privacy/security design for V2X systems.

  4. [4]

    K. Rathod et al.. Comparative Analysis: LLM Families on Legal Benchmarks.” Internal Technical Report, CourtEasy.ai / Nugen, 2025.

    InLegalBERT, InLegalLLaMA, and GPT-4o-mini evaluated on LegalBench and NyayaAnumana, synthesizing 15+ papers to inform production RAG workflow design and model routing.

3.2 Interactive marginalia

acts = model.trace("systems that retrieve, reason & remember.")
plot_attention(acts, layer=11, head=4)

Fig. 3.2 — The attention motifs this section writes about, run over the site’s own thesis. Synthetic weights, real mechanisms — previous-token, subject-binding & semantic-recall heads. Executable — press ▶ to re-run.

A

Appendix: Proficiencies

Table 1. Technical proficiencies, grouped by domain.

DomainStack
LanguagesPython · TypeScript / JavaScript · Java · C++ · Swift · Rust · SQL
AI / MLRAG & GraphRAG · LLMs & transformers · multi-agent systems · LangChain / LlamaIndex · PyTorch · vLLM
Cloud & InfraAWS (Bedrock, Lambda, S3) · Azure · Docker / Kubernetes · event-driven microservices · SLURM / HPC
Web & MobileReact / Next.js · SwiftUI · Tauri · Flask / FastAPI · Tailwind CSS
DataPostgreSQL / MySQL · DynamoDB · NebulaGraph · ChromaDB / Pinecone · PySpark
ToolingGit · CI/CD / Vercel · LaTeX · MCP servers
Plate A
T(2,3) · torus knot↻ drag to rotate
Plate A — Proficiency manifold, rendered as a T(2,3) torus knot. Vertices mark sampled skills; topology is, of course, a joke. Interactive — drag to rotate.

Correspondence

Open to research collaborations and engineering problems worth losing sleep over.

Ω

Open questions

end paper — the questions behind the preceding pages