Open to work MS CS (AI Track) at Utah, graduating May 2027. Looking for AI / SWE internships and new-grad roles. Get in touch
Software engineer · Applied AI · Salt Lake City

Point me at anything and give me enough time.

I'm a versatile engineer, not a one-stack specialist. Five years of production work across backend, infrastructure and full-stack, then a master's in AI, and the pattern has always been the same: hand me an unfamiliar domain, and I'll learn it well enough to build the real thing.

Enterprise backup, healthcare telephony, atmospheric science, neuroimaging, reinforcement learning, retrieval systems. What I keep coming back to is AI applied to somewhere it hasn't been yet, and that's the part I find genuinely fun.

See what I've built → GitHub LinkedIn
Portrait of Hitesh Kumar Balegar
MS CS, AI TrackGPA 3.8
Industry5 yrs
LabsHolmes · Merenstein
If you're short
on time
01

Selected work

Click a card to open the case study
Filter
02

Experience

Five years industry · two research labs · two TA roles
Industry · BETSOL, Bangalore
Aug 2020
Jun 2025
4 yr 11 mo

Full Stack Software Engineer · Zmanda Enterprise

Enterprise backup platform · Go · Python · Kubernetes
  • Zero-touch provisioning across 470+ locations. Go CLI tooling and Python automation packaged as AWS Lambdas, with idempotent environment-aware config pipelines. Rollout time 6 weeks → 48 hours.
  • Monolith to microservices. Re-architected a legacy backup platform into Docker services on Kubernetes with rolling updates, health probes and namespace-scoped isolation. New-environment setup −70%, environment drift eliminated.
  • Backend performance. Profiled Django ORM queries with EXPLAIN ANALYZE, added connection pooling, query-result caching and multithreaded WSGI workers. p95 API latency −60% under heavy backup/restore load.
  • Security & compliance. FIPS-compliant encryption modules and SELinux MAC packages across the stack; HashiCorp Vault for dynamic secret injection, removing every hardcoded credential across 10+ TB of customer data.
  • RAG support chatbot. LangChain + GPT-4 over Qdrant and Pinecone, with an automated RAGAS evaluation pipeline tracking faithfulness, context precision and answer relevancy. Software quality practice applied to a generative feature.
GoPython / DjangoKubernetesAWS LambdaVaultLangChainQdrantRAGAS
Sep 2020
Jun 2025
client engagement

Full Stack Software Engineer · CVS Health & Aetna

Fortune 10 healthcare · telephony analytics · HIPAA
  • Telephony analytics portal at scale. FastAPI + React on Azure Kubernetes Service processing 5M+ call records a month at 99.9% uptime, with RBAC, field-level encryption and structured audit logging on every data path.
  • Cross-system integration. RESTful layers spanning ServiceNow, telecom provider APIs and internal warehouses, feeding real-time reporting used daily by 50+ analysts and leadership.
  • Call-routing re-platform. Led the move of legacy routing logic into Node.js/Express microservices with circuit breakers, structured logging and OpenAPI contracts. Higher throughput, measurably fewer on-call incidents.
  • Forecasting for the contact center. Custom Erlang B extended traffic models and ML demand pipelines surfaced in Power BI and ServiceNow. Forecast accuracy +25%, staffing allocation +20%, fewer SLA breaches.
  • Release engineering. Helm chart versioning, Jenkins multi-branch pipelines and GitHub Actions with smoke tests and rollback triggers. Release cycles −40%, deployment setup from a full day to under 2 hours.
FastAPIReactAzure AKSNode.jsHelmJenkinsPower BI
Research & teaching · University of Utah
Jan 2026
present

Research Engineer & Graduate Mentor · Holmes Lab

  • Biomass burning in Nepal. Compared FINN, GFED and QFED inventories and fused MODIS polar-orbiting with GEO-KOMPSAT-2A geostationary FRP into an hourly emissions framework feeding HYSPLIT dispersion.
  • Dust exposure along the Wasatch Front. Integrated air-quality observations, source apportionment, satellite retrievals and population estimates into exposure figures delivered to epidemiologists at UNR.
  • Pipelines & mentoring. Co-built openaq-monitors-extractor for global PM2.5/AOD aggregation; mentoring HAPPIEST undergraduates on ML workflows and scientific computing.
xarrayGeoPandasNetCDFXGBoost
Jan 2026
present

Teaching Assistant · CS 6017 Data Analytics & AI for Business

  • Supporting a graduate course on supervised learning, clustering, regression and visual representation, helping students build analytical pipelines and dashboards end to end.
  • Built course-ready applied-AI demos for the business school: a LangGraph voice agent, an SEC-filing RAG research agent, browser automation utilities and fine-tuned LLMs.
Jan 2026
May 2026

Research Assistant · Merenstein Lab

  • Extraction and fusion pipelines over OpenAQ, EPA, PurpleAir, MODIS and VIIRS supporting air-pollution forecasting and demographic analysis.
  • Neuroimaging processing and analysis pipelines for cognitive neuroscience across large, heterogeneous datasets.
03

Stack

Things I've actually shipped with
Languages
PythonGoTypeScriptSQLJavaC#
ML & AI
PyTorchTransformersPEFT / QLoRALangGraphLlamaIndexRAGASFAISSQdrantGymnasium
Data & distributed
Apache SparkSpark MLlibKafkaPostgreSQLCHPC / HPCxarray
Platform & DevOps
DockerKubernetesHelmGitHub ActionsTerraformAWSAzure AKSNext.js
04

Writing & publication

Essays and case studies
IJRPR
2023

Analyzing Multilingual LLMs Using Pre-Trained Dataset Model

Benchmarked mBERT, XLM-R and GPT across cross-lingual transfer, sentiment analysis and generation using BLEU, ROUGE, F1 and perplexity over diverse language families. Vol. 4, No. 11, pp. 97–105.

Paper ↗
05

Recognition & education

Awards
20261st place · Game Day Analytics Challenge · 120+ students, 59 brands
2026Honorable mention · Generative AI Hackathon · "Most deployable, tech-driven solution", 35+ teams
20251st place · Salesforce IDEAthon · recruiting CRM for Utah Athletics
21–24Team Excellence, Above & Beyond, Bug Bounty · BETSOL
Education

University of Utah

2025–27
MS Computer Science, AI Track · GPA 3.8
Multimodal LLMs · Advanced AI · Managing Data for ML
06

Commit activity

— contributions · last 12 months
—
Less
More
—
longest streak
—
active days
10
public repos
Python · Go
most used