Building backend and data platforms at Aztlan

I’m Jagdeep — a backend engineer. I build data pipelines, AI platforms and the infrastructure they run on.

Since 2024 I’ve shipped a Rust log pipeline moving 20,000+ events a second, an AI backend that reads identity and financial documents, a compliance-governance platform, and the bare-metal running a dozen internal services — at Aztlan now, at Sequoia Consulting Group before that. Rust and Python mostly, with Go and Node where they fit better.

Experience

2025 —

Backend Developer · Aztlan

Log infrastructure, document intelligence and AI governance — plus the bare-metal that runs them. Hyderabad.

2024 — 25

Backend Engineering Intern · Sequoia Consulting Group

Compensation tooling and notification infrastructure for 20+ client teams. Bangalore.

2021 — 25

B.E. Computer Science Engineering · Chitkara University

Graduated with a CGPA of 9.72.

Selected work

LogTrim

AZTLAN · 2025

An end-to-end log processing pipeline written in Rust.

  • Handles 20,000+ events per second across five integrated applications.
  • Cut Splunk ingestion costs by roughly 40% by deduplicating and aggregating in real time, before anything is forwarded.
  • Config reloads with zero downtime, using PostgreSQL LISTEN/NOTIFY and thread-safe memory swapping.

Rust · Tokio · Axum · Fluvio · Kubernetes

UWS — Document Intelligence

AZTLAN · CLIENT PROJECT

An AI backend that pulls structured data out of identity and financial documents.

  • Extracts fields from Aadhaar, PAN, US IDs, payslips and bank statements.
  • A policy engine checks the results against compliance rules and flags cross-document mismatches — a salary that disagrees between a payslip and a bank statement, for instance.

Python · FastAPI · PostgreSQL · GCS · Gemini API

Core V1 — AI Governance

AZTLAN · 2025

A platform for enforcing compliance policy across AI models and datasets.

  • Teams define NIST and ISO policies, then group models and datasets into governed use cases.
  • Each use case is evaluated through policy questionnaires and against uploaded compliance documents.

Python · FastAPI · SurrealDB

Infrastructure & Internal Tooling

AZTLAN · ONGOING

The bare-metal that runs the company’s internal services, and everything on it.

  • 10+ services deployed and maintained — GlitchTip, Uptime Kuma, Outline, Docmost, Postiz, Papra, Unleash.
  • Reachable securely from outside through Cloudflare Tunnels, with monitoring and issue resolution.

Proxmox · Coolify · Cloudflare · Authentik · Docker

Compensation Tooling

SEQUOIA · 2024 — 25

Merit-cycle tooling and notification infrastructure for 20+ client teams.

  • A merit-plan cloning feature cut cycle setup from hours to minutes.
  • A notification service on SNS, SQS, SES and Lambda delivers 1,000+ emails a month at 99% reliability.

Python · Go · MongoDB · Redis · AWS

A pipeline that converts uploaded video into HLS for adaptive bitrate streaming.

  • Transcode jobs fan out to ECS Fargate through SQS and EventBridge.
  • Nothing is kept warm between uploads — the cost is only what is actually transcoded.

Node.js · AWS ECS · FFmpeg · MongoDB

Toolkit

Languages

Rust, Python, Go, JavaScript, SQL

Frameworks & data

FastAPI, Axum, Tokio, Celery, PostgreSQL, MongoDB, Redis, SurrealDB

Infrastructure

Kubernetes, Docker, AWS, GCP, Proxmox, Coolify, Cloudflare

Tooling

Fluvio, Fluent Bit, OPA Gatekeeper, Authentik SSO, Git, Linux

Get in touch

Backend and infrastructure work, Rust or Python, or an interesting scale problem — those are the emails I answer fastest.