Drawn by A. Chaturvedi
Discipline Backend / Distributed Systems
Location Kanpur, IN
Time --:--:--

Sheet 01 — Elevation

Ashutosh
Chaturvedi

Backend & distributed systems engineerdrafted before it's built.

3rd‑year B.Tech CS student at PSIT Kanpur, working toward Intelligent Systems Engineer and Infrastructure Architect roles. I spec a system on paper, build it, watch it fail under load, and correct the drawing. Async pipelines, real‑time platforms, and a forecasting model reading solar flares off ISRO's Aditya‑L1 satellite all went through the same loop.

LeetCode
400+
Status
3 shipped, 1 in build
Institution
PSIT Kanpur
Availability
Open — internships

Sheet 02 — Work

What's been built

Fig. 2.1 IN BUILD

Knoxis

Degree mini-project

A solar flare prediction system built directly on ISRO's Aditya‑L1 satellite data — reading raw instrument output rather than a cleaned public dataset.

  • Python pipeline downloads, audits, and consolidates SoLEXS/HEL1OS data from ISRO's PRADAN portal
  • Random Forest classifier forecasting flare activity across a 6‑hour window
  • Planned: a Next.js + FastAPI dashboard replaying flare events in simulated real time
Pythonscikit-learnFastAPINext.js
Repo ↗
Fig. 2.2 SHIPPED

Developer Notification Platform

Flagship build — 5 phases

A modular-monolith notification system built phase by phase, from multi-tenant auth to a gRPC layer shared across sync and async paths.

  • JWT multi-tenancy with HMAC-verified webhook ingestion
  • RabbitMQ pipeline with a dead-letter queue and async email delivery
  • STOMP/WebSocket push backed by a live session registry
  • Redis caching and Elasticsearch full-text search
  • gRPC + Protobuf sharing business logic across both delivery paths
JavaSpring BootPostgreSQLRabbitMQRedisElasticsearchgRPC
Repo ↗
Fig. 2.3 SHIPPED

Convo

Live in production

A real-time chat application running in production, not just deployed once and left alone.

  • WebSocket fan-out with Redis TTL-based presence and delivery receipts
  • JWT access/refresh auth, group rooms with join codes, direct messaging
  • Fixed a stale-closure bug in the socket layer and a public Redis port caught via a hosting alert
FastAPIPostgreSQLRedisNext.jsDockerNginx
Live site ↗
Fig. 2.4 SHIPPED

Habit Tracker

First full-stack build

The project that started it all — learning the full stack end to end by shipping something real instead of following a course.

  • FastAPI backend, PostgreSQL storage, React + Tailwind frontend
  • Deployed on Render — first taste of what happens after the code works locally
FastAPIPostgreSQLReactTailwind
Repo ↗

Sheet 03 — Specification

What I'm pointed at

Distributed Systems
Queues, consensus, and the failure modes that only surface under load.
AI Infrastructure
The plumbing underneath models — serving, pipelines, the data that feeds them.
Cybersecurity
Thinking about a system the way someone trying to break it would.
Autonomous Agents
Software that has to decide, not just respond to a request.
Cloud & Databases
Where data actually lives, and what it costs to keep it consistent.
Networking
The layer everything above quietly depends on.
Toolbox
  • Java / Spring Boot
  • Python / FastAPI
  • PostgreSQL
  • Redis
  • RabbitMQ
  • gRPC / Protobuf
  • Elasticsearch
  • Docker
  • React / Next.js
  • WebSockets
  • Linux

Sheet 04 — Revision log

Currently

Sheet 05 — Sign-off

Let's talk systems.

Open to internships, collaborations, and anyone who wants to argue about queue design.