SOFTWARE ENGINEER — BACKEND & SYSTEMS

Satva Shah

Backend-leaning CS student who likes taking systems apart to see why they hold together — async messaging, multi-agent pipelines, and the CI/CD scaffolding that keeps it all shipping.

satva@systems:~$  align curiosity ↔ deliverability
GitHub ↗ LinkedIn ↗
01 — About

I'm a third-year CS student who ends up on the backend by instinct — I'd rather understand how a system fails before I decide how to build it. Most of my work sits somewhere between distributed systems and applied AI: producer-consumer queues, multi-agent orchestration, RAG pipelines, and the pipelines that deploy them. I tend to say yes to things I haven't done before and figure out the rest on the way.

02 — Experience
Software Engineering Intern
Security Services Company · Pune
  • Phase 1Built an asynchronous producer-consumer workflow in Go with RabbitMQ for real-time messaging between services.
  • Phase 2Built an async multi-agent orchestration layer in FastAPI connected to RAG pipelines, including a multilingual AI agent that translated ingredients across regional Indian languages for a food platform.
  • CI/CDTook on CI/CD for an 8-module microservices repo with no prior experience — GitHub Actions, Docker, a matrix build strategy, and dual image tagging (latest + commit SHA) that wasn't asked for.
India AI Impact Summit
Top 200 of 38,000 teams
  • Built a multi-agent system to detect and intercept digital scams targeting senior citizens with low digital literacy.
  • Independent research surfaced service-based phishing — fake repair calls, fraudulent agents — a pattern underreported in mainstream data but confirmed by NGOs as the most common real-world case.
  • Presented live to a non-technical panel, reframing the pitch around human outcomes rather than architecture.
03 — Selected Work
FLUTTER · DART
Khao
A food-focused mobile app, built alongside my work on multilingual ingredient translation.
 JAVASCRIPT
claude-token-counter
A small utility for estimating token usage in Claude prompts and conversations.
 DATA / ETL
ETL_F1
An ETL pipeline built around Formula 1 race data — extraction, transformation, and reporting.
 DBT · POWERSHELL
DBT_AirBnb
A data modeling project applying dbt to Airbnb listings data.
04 — Beyond Code
→
Synapse, ML Committee — Mentored juniors in machine learning, moving from being mentored to mentoring within the same semester. Learned that the job is usually to diagnose confusion, not add more explanation.
→
SBMC, Social Work Committee — Regular visits to Atam Memorial Hospital, spending time with children being treated for cancer. Not a line item — just something I show up for.