Nagendra
Chaudhary

AI & full-stack engineer. I build agents, RAG systems, and the eval harnesses that prove they work. Currently Founding Engineer @ AltOps.

500+ users - on Quipler~$10.3K - TCU CREATE funding for Quipler800ms -> 50ms - latency cut (~94%) at Brazos Safety Systems3-5 days -> 30min - document processing automated at Boys & Girls Clubs of America
scroll - the 8-second version is next
// the 8-second version

Keep scrolling. This is the short version.

01 / 10
500+ users

on Quipler, his campus marketplace - live at quipler.app

LIVE
02 / 10
~$10.3K

TCU CREATE funding for Quipler

LIVE
03 / 10
800ms -> 50ms

latency cut (~94%) at Brazos Safety Systems, 99%+ availability

EVAL
04 / 10
3-5 days -> 30min

document processing automated at Boys & Girls Clubs of America

EVAL
05 / 10
4.8s -> 730ms

query speedup over 500K+ records at Market Match

EVAL
06 / 10
5,000+ patients

RAG over clinical data across 100+ trials at Pathi Health

EVAL
07 / 10
82.6% -> 91.3%

critical-field accuracy on ProxyEval after fixing a real failure

EVAL
08 / 10
30/30 tests

ResolveDesk on real Postgres, incl. crash mid-refund

EVAL
09 / 10
AAAI 2026

co-authored AI4EDU workshop paper, ~40% hallucination cut

AWARD
10 / 10
3rd place

Murf Falcon Buildathon - AudioComics

AWARD
that was 8 seconds.

The long version is below - every project, labeled real vs demo.

// proof of work

Everything he has built

Every project is labeled plainly: live product, measured eval, award, demo, or work in progress. No invented numbers.

proof of work

live products, real users, measured results

Quipler

500+ USERS

Campus marketplace with 500+ users and ~$10.3K in TCU CREATE funding. Live product - try it.

500+ real users, ~$10.3K funding, in production

Next.js / Supabase / Stripeopen ->

ResolveDesk

TESTED

AI support platform: agent answers tickets from policy docs with citations, auto-resolves safe ones, routes refunds. Crash-safe actions, duplicate-event protection.

30/30 tests on real Postgres (incl. crash mid-refund). Eval on 24 hand-written tickets: routing 100%, citation match 91%. Load: 500 tickets, 61/s, p95 577ms. Tickets are synthetic - README says so.

Python / TypeScript / Postgres / Dockeropen ->

ProxyEval

REAL FILINGS

Extracts governance data from real SEC DEF 14A proxy statements - but the evaluation harness is the point: gold set, baseline, one real failure found and fixed, regression test.

8 real SEC filings, 31 hand-checked gold labels. Critical-field accuracy 82.6% -> 91.3% after fixing a real parsing failure (now a regression test).

TypeScript / eval harnessopen ->

FasterFlow

REAL PRODUCT

Workflow-automation product. Founding engineer (part-time) - core features shipped to real users.

Live product with real users (private).

TypeScript

more builds

hackathons, awards, targeted engineering

AudioComics

3RD PLACE

Turns comics into audio in 10+ languages (OpenAI Vision + Murf AI). 3rd place, Murf Falcon Buildathon.

Placed 3rd at the Murf Falcon Buildathon.

Python / Vision / TTSopen ->

Blackbox

SYNTHETIC DATA

AI on-call incident investigator: builds an evidence-cited timeline from outage logs, ranks root-cause hypotheses, drafts the postmortem.

50 tests + eval harness with honest scores. Runs on synthetic incidents - README says so.

FastAPI / Reactopen ->

sel-sensor-hub

141 CHECKS

C11/C++17 sensor data pipeline: ring queue, CRC-16 packet protocol, fixed-point filters, lock-free SPSC queue.

141 unit checks + 100k-case deterministic fuzzing under ASan/UBSan. Built to mirror SEL's real firmware job posting.

C / C++open ->

SafeMap

LIVE DATA

San Francisco safety map - live crime + theft reports to find safe parking.

Working app against live public data.

TypeScript / Mapsopen ->

in progress & experiments

labeled honestly

Wisper

IN PROGRESS

His own agentic automation engine: durable tasks, approval-gated actions, audit trail, connectors. Gmail connector is live (real OAuth); other connectors are tested mocks, labeled as such.

Live Gmail integration; remaining connectors are mocked adapters - stated in the repo.

TypeScript / Temporal / Postgresopen ->

Signal Fire

HACKATHON

Privacy-first people discovery: scan a room QR, post one need and one offer; a match fires only on mutual trade. Field-level consent enforced in code. Universal Hackathon entry, Oct 2026.

Matching, expiry, consent and revocation implemented and locally tested; agents are offline mocks - README lists exactly what is and isn't built.

Python / agents

The-Swarm

IN PROGRESS

Hyperlocal anonymous whisper map - posts visible within a 20-minute walk, trending by reactions.

Active build, core map + reactions working.

Flutter / Supabaseopen ->

Older projects

ARCHIVE

Umoja, ERS-Assistant, SpaceMed AI, TensorFrames, Hogwarts Artifact Online, Project Controls Dashboard (Power BI/DAX).

College-era builds, kept public on GitHub.

variousopen ->
PUBLISHED

AAAI 2026 - AI4EDU Workshop (co-author)

Multi-agent adaptive tutoring system from the TCU NextGen AI Lab.

RAG over 1,000+ academic documents - ~40% fewer hallucinations - ~70% less instructional-design effort

LEADERSHIP
  • - President, HACKTCU (2025-26)
  • - Vice President, Robotics & AI Club (2025-26)
  • - President, Venture Builder TCU (2025-26)
  • - HackMIT semifinalist - accepted to HackMIT, HackHarvard, Cal Hacks, TreeHacks
// timeline

Where he has worked

AltOps

Founding Engineercurrent
  • Building computer-use agents that automate real work end to end

Brazos Safety Systems

Software EngineerAug 2025 - May 2026
  • Cut core latency 800ms -> 50ms (~94%)
  • Held 99%+ availability in production

FasterFlow

Founding Engineer (part-time)
  • Core engineer on a live workflow-automation product with real users

Pathi Health (Vaccinity)

AI Engineer Intern2024
  • RAG pipelines over 5,000+ patient records across 100+ clinical trials

Boys & Girls Clubs of America

Software Engineer
  • Automated financial-document processing: 3-5 days -> ~30 min (Python, OpenCV, Gemini -> Salesforce)

Market Match

Software Engineer Intern
  • Sped up queries over 500K+ records from 4.8s -> 730ms
// the person

From Nepal to TCU

Nagendra is from Nepal. He came to the United States for his studies and is completing a B.S. in Computer Science at Texas Christian University.

Since then he has not stopped building: founding-engineer roles, products with real users, and published AI research.

B.S. Computer Science (minor Mathematics), Texas Christian University, May 2026. Dean's List 2024 & 2025, Next Genius Full-Tuition Scholarship, CodePath AI Engineering Fellow.

// off the keyboard
loves to travelplays soccerplays ping-pongworks outmeets new people everywhere he goeslearns new cultures
FOCUS
AI agentsRAG pipelinesevaluation harnessesworkflow automationcomputer-use
STACK
PythonTypeScriptJavaScriptSQLC/C++FastAPIReactNext.jsPostgreSQLTemporalDocker
// ask me anything

The concierge

Grounded only in verified facts about Nagendra. If it does not know, it says so - it never invents.

Ask me anything about Nagendra - his work, numbers, story. I only answer from verified facts, and I say so when I do not know.