Jeethesh avatar

Hi, I'm Jeethesh,

Software / AI Engineer
who ships distributed systems with
Evals + Grounding + Ops

My path

Work experience

End-to-end ownership across discovery → delivery → adoption, collaborating closely with stakeholders and engineering.

Customer data publishing

7 engineering data types to a centralized platform for downstream apps and calculations

AI-assisted quality engineering

Dual-agent Stryker loops, 90% mutation-score target, 50%+ less testing effort across 10+ repos

Kafka platform reliability

Cleared 200+ stale topics and automated age-based cleanup for 10+ teams

Event-driven orchestration

Node.js, Kafka, Redis, WebSockets, and Angular for well-engineering workflows

External data ingestion

Pipeline for 4 third-party data products into internal engineering calculations

Compliance & observability

New Relic tracing, structured logs, activity audits, and technical-status notifications

Halliburton

Software Engineer

July 2024 – Present

Houston, TX

  1. Built a customer-facing data publishing feature using Node.js, Kafka, Redis, WebSockets, and Angular, enabling drilling and well engineers to publish 7 types of engineering data to a centralized platform for use by downstream applications and engineering calculations.

  2. Built a dual-agent testing framework that detects repository test setups, generates unit tests, and iteratively improves them using Stryker mutation feedback to reach a 90% mutation-score target, reducing overall testing effort by 50%+ across 10+ repositories.

  3. Prevented recurring Kafka capacity issues across 10+ teams by clearing 200+ stale topics and automating age- and inactivity-based cleanup to protect dependent services.

  4. Automated test quality validation in GitLab CI across 10+ repositories using mutation testing, refreshing baselines after merges and preventing manual edits so developers could validate code changes without maintaining testing artifacts.

  5. Enabled new customer features by building a data retrieval pipeline that ingests 4 external company data products from a central domain into internal systems, unlocking access to third-party data for engineering calculations.

  6. Integrated New Relic, structured logging, activity auditing, and Kafka technical status notifications for end-to-end observability.

What I build

Projects

A retrieval-augmented assistant over personal resilience notes — Flask orchestrates Gemini + Pinecone with Cohere reranking, citation-grounded answers, Firebase auth, and Cloud Run deployment.

Why I built it

I wanted to stay consistent through personal challenges, but generic habit trackers never felt personal enough to stick. Building one around my own reflection process meant I could shape it to how I actually think — and feel invested enough to keep using it.

  • 30/30 golden CI tests and 6/6 live RAGAS evals (1.0 faithfulness, 0.90 relevancy)
  • Intent-aware routing that bypasses vector search for aggregate queries
  • Langfuse per-stage tracing for retrieval and generation failures
  • Per-user rate limiting and server-side secret management on Cloud Run

30/30

Golden tests

1.0

Faithfulness

0.90

Relevancy

Stack

GeminiPineconeCohereFlaskFirebaseCloud RunLangfuseRAGASReact

Education

Academic foundation

Data science graduate paired with a strong engineering undergraduate core.

University of North Texas

M.S., Data Science

Aug 2022 – May 2024 · GPA 3.9/4.0

National Institute of Technology Jamshedpur

B.Tech., Electronics & Communication Engineering

July 2018 – May 2022 · GPA 7.6/10.0

Capabilities

Skills

Systems, AI, cloud, and backend craft — grouped by how they show up in shipped work.

Languages

EnglishProfessional proficiency
TeluguNative
HindiProfessional proficiency
SpanishStill buffering…

Distributed Systems

  • Node.js microservices
  • Kafka / Confluent
  • Redis
  • WebSockets / Socket.IO
  • Idempotency & job state
  • Observability (New Relic, Sentry)

AI / LLM

  • RAG & vector search
  • Embeddings & reranking
  • LLM evaluation (RAGAS)
  • Prompt engineering
  • Observability (Langfuse)
  • Structured outputs
  • AI agents & mutation testing
  • Computer-use agents
  • Local LLM inference

Cloud & DevOps

  • Google Cloud Run
  • Firebase
  • Docker / Kubernetes / Helm
  • Azure DevOps
  • GitHub Actions / Jenkins
  • CI/CD

Backend & Languages

  • Python
  • TypeScript / JavaScript
  • C# / Java / SQL
  • Flask / FastAPI
  • Express / Hapi.js
  • Angular / Next.js / React
  • PostgreSQL / Supabase / Pinecone / SQLServer

Credentials

Certifications

Signals of continued learning across AI tooling and modern developer workflows.

Ask me

Chat with a grounded copy of my profile

Ask about my experience. Speak or type — answers stay scoped to this portfolio.