$ *^//| /!>|<@<
$*%$ >/#<^!
CS @ University of Utah. Distributed systems, agentic AI, full-stack. I ship things that run in production.
selected work
DataForge
Built governed AI platform generating Databricks apps from natural language in under 60s. 3-agent LLM system: SQL Generator, Self-Healing Agent (3-retry), async Insight Generator. Hackathon Winner, 1st/50+ teams.
SLO-Guard
Provisioned isolated Kubernetes namespaces per PR, ran k6 load tests, gated merges on SLO thresholds. Under 3-min MTTR via Grafana dashboards and Alertmanager.
StreamEval
Distributed C++/Python pipeline continuously running parameterized scenario tests, scoring software behavior against safety thresholds across distributed worker nodes.
more projects
Music Recommendation System
Full-stack recommendation engine processing 10,000+ Spotify tracks with cosine similarity. Top-10 results in under 200ms.
AthleteIQ
Production-quality Android app surfacing live game stats via multi-screen, multi-module Kotlin architecture with reactive state management.
API Gateway Service
High-throughput gateway with JWT auth, rate limiting, Redis caching. 150ms response improvement. 500 concurrent requests sustained under load testing.
I'm a CS junior at the University of Utah obsessed with distributed systems and agentic AI. I build things that run at scale — from multi-agent LLM pipelines to Kubernetes infrastructure. Currently doing AI research and network engineering at Utah while looking for internships where I can ship hard things.
AI Research Assistant
University of Utah · May 2025–Present
- – Context protocols for 15+ node multi-agent LLM systems
- – Documented 12 attack surfaces with mitigation strategies
Network Ops Engineer
University of Utah · Sep 2025–Present
- – Async pipeline polling 2,000+ switches via SNMP
- – Endpoint lookup cut to under 5 seconds via Nautobot
ML Intern
Garje Marathi · May–Oct 2024
- – Knowledge Graphs into RAG system, +35% accuracy
- – Graph load time cut 90% — 33hrs to 3hrs for 1M+ nodes
Languages
AI + ML
Systems
Cloud