// AI ENGINEER · DATA ANALYST

VATSAL DOSHI

I build LLM/RAG systems — and the analytics that prove they work. Currently shipping recommendations to 50K+ daily queries as AI Engineer at Connyct.

CHICAGO, IL 41.88°N / 87.63°W SCROLL
// SEPARATES THE CINEPHILES FROM THE "I WATCH STUFF" CROWD
01 PROOF, NOT ADJECTIVES
0

DAILY QUERIES SERVED BY A REC ENGINE I TUNE AT CONNYCT

0

CITATION ACCURACY ON A RAG SYSTEM DELIVERED TO ELASTIC — UP FROM 90%

0

PARTNER DOC SEARCH TIME — DOWN FROM 10+ MINUTES

0

MANUAL REPORTING TIME CUT FOR A DC NONPROFIT SERVING 1,500+ STUDENTS

02 ABOUT ME

The AI proposes.
Deterministic code executes.
And every number gets a defense.

I'm Vatsal — an engineer from India, via an M.S. at the University of Maryland, now working remotely from Chicago. Days go to Connyct, where I run the AI features of a short-video app for college students; nights go to building tools like Switchboard and Marquee — a movies + TV tracker I built because I log everything I watch.

Off the clock: Formula 1 race weekends, planning trips around roller coasters (the enthusiast kind of enthusiast), and a boba habit I refuse to quantify — the one number that doesn't get a dashboard.

I work the full loop: user interviews → SQL/pandas analysis → LLM systems in production → dashboards that tell leadership the truth. My projects share one rule — models are never trusted with what code can verify. Retrieval gets test suites, recommendations cite their sources, and migrations ship with rollback plans.

DATA & ANALYTICS
SQLPYTHON / PANDASTABLEAUPOWER BIA/B TESTINGREGRESSIONGCP / BIGQUERY
AI / LLM
RAGLLM EVALSPROMPT ENGINEERINGLANGCHAIN / LANGGRAPHCHROMADBHYBRID SEARCHSTRUCTURED OUTPUTSFINE-TUNING (LORA/PEFT)MCP / COMPUTER-USE AGENTSREC SYSTEMS
ENGINEERING
FLASK / FASTAPIREACTDOCKERCI/CDGIT
SKILLS, PHYSICS-ENABLED — DRAG THEM
03 EXPERIENCE
2026 —
NOW

AI Engineer @ Connyct

Short-video social platform for college students, with a Warner Music Group content partnership. Scaled the content recommendation engine to 50K+ daily queries; 92% user satisfaction across AI features from 50+ user interviews.

2025

AI Engineer @ Elastic — consulting engagement, UMD

Built and delivered the Partner-Docs RAG assistant below — hybrid retrieval, 50–60% latency cut, a 25-case accuracy suite, and iterative delivery with Elastic client stakeholders.

2025

Data Analyst Intern @ Tech Turn Up

Sole data analyst for a DC nonprofit teaching 1,500+ K-12 students. Built the analytics stack from scratch — 8 Tableau dashboard pipelines, 75% less manual reporting, and cohort analysis that drove an executive restructuring.

2024 —
2025

M.S. Information Systems @ University of Maryland

College Park, MD. Coursework across data analytics, database management, designing AI systems, and cloud computing & big data. Certified ScrumMaster; Google Project Management Certificate.

2020 —
2024

B.E. Computer Engineering @ Gujarat Technological University

Ahmedabad, India — CGPA 3.8/4.0.

04 SELECTED WORK
05 RESUME

One résumé,
compiled to target.

Same career, same numbers — weighted for the role you're actually hiring for. Pick a target and it builds. No email gate, no form.

REV 2026.07 · 1 PAGE · EVERY CLAIM SOURCED

// SELECT COMPILE TARGET

06 CONTACT

OPEN TO DATA / AI ANALYST & RAG-LLM ENGINEER ROLES

prefer a CLI? press [/] and type 'help'. there are secrets.