Skip to content
Open to teams / software engineering internships

Srinesh Rudraksha

Building production-grade AI & full-stack systems.

I build and measure real systems — Docker services under load, React apps with auth, and Python data/ML tools. I care more about working code and tests than demos.

Hyderabad, India (IST)B.Tech CS · MVSRECGraduating 2027
Products shipped
06
Parser tests · Spentd
90+
Android build · v1.1.5
APK
Docker · Traefik lab
Load
01

Selected work

Every project below is deployed or released. Links go to the live build and the source.

01Expense tracker · Android + API

Spentd

Live site · APK v1.1.5 · ~90 parser tests

Privacy-first Android expense tracker. It reads payment screenshots with on-device OCR and parses bank/UPI SMS; screenshots stay on the phone and only structured fields sync.

Monorepo with an Expo app, a Hono + Postgres API, and shared Zod parsers covered by ~90 unit tests. Username + passcode auth (Argon2id on the server, JWT client session), budgets and cash wallets, CI and APK releases through v1.1.5.

  • Expo
  • React Native
  • Hono
  • Postgres
  • Drizzle
  • Zod
  • OCR + SMS parsers
  • GitHub Actions / EAS
02Distributed systems lab

ReadIssue

50 pytest cases · live topology · chaos scenarios

Overload simulation lab: real Docker workers behind a Traefik load balancer (worker cap 8), ~400k seeded product rows, a FastAPI control plane, and a live React dashboard with autoscaling and chaos scenarios.

Measured on the read path: indexed queries ~38 ms p95; with PgBouncer, Postgres backends for 4 workers fell from ~42 to ~10; the replica path cut primary multi-core CPU (Docker stats) from ~550% to ~1%.

  • FastAPI
  • Docker
  • Traefik
  • PostgreSQL
  • Redis
  • React
  • WebSockets
03AI calendar

Kairo

73 unit tests · LLM + RAG event pipeline

AI calendar that turns natural language, images and PDFs into structured events through a Groq LLM + RAG pipeline.

Supabase handles auth, sync and realtime; a WhatsApp bridge (QR intake) allows event creation from chat. 73 unit tests caught 3 pre-production issues.

  • React
  • Vite
  • TypeScript
  • Zustand
  • Supabase
  • Groq
  • Vitest
  • WhatsApp bridge
Also shipped
UML whiteboard04

Boards

UML whiteboard on React Flow — custom class, actor and use-case nodes, relationship edges, local-first IndexedDB persistence, PNG/PDF export. AI diagram generation runs through a Worker proxy, so no API keys sit in the client.

Local-first · PNG/PDF export

  • React
  • TypeScript
  • React Flow
  • Dexie / IndexedDB
Logo super-resolution05

pngmodel

Fine-tuned Real-ESRGAN that upscales low-quality logos into high-res PNGs and clean SVG vectors. Two-stage GAN training (L1 → perceptual + GAN) on an RTX 3070 Ti over ~1,860 images, served through a Gradio UI.

x4 super-res · PNG + SVG

  • Python
  • PyTorch
  • Real-ESRGAN
  • Gradio
Rust TUI search launcher06

ddg-spotlight

Spotlight-style terminal search launcher for Hyprland: live, debounced DuckDuckGo instant answers in a floating TUI, with runtime theming and browser hand-off.

~90k paths in ~0.1 s

  • Rust
  • ratatui
  • crossterm
  • reqwest
02

Experience

Jun 2025

AI & Data Science Intern

OTBI · Osmania University

Hyderabad

DomainGroundwater / borewell mapping from field resistivity data

  • Built a Python pipeline that turns multi-depth resistivity CSVs into contour maps, denoised profiles and frequency graphs.
  • Added a drilling-point predictor so site picks are not only visual guesses from charts.
  • Shipped a Tkinter GUI for non-technical users to upload data, set depths and export maps — compared in-lab against commercial-style instruments such as the PQWT GT500.
  • Python
  • Pandas
  • NumPy
  • SciPy
  • Matplotlib
  • Tkinter
03

About

I'm a third-year Computer Science student at MVSREC in Hyderabad. Most of what I know came from building things end to end — Docker-based services, React apps with auth, Python data and ML tools — then checking whether they actually hold up.

The problems I like span the stack. One week it's overload paths and load balancers; the next it's an LLM pipeline turning PDFs into structured events. I'd rather own a problem through design, implementation and tests than hand over a demo.

Curious by default: I read more code than I write. Open to teams shipping real product, including internships where I can contribute to a production codebase and learn from senior engineers.

Education

B.Tech, Computer Science & Engineering

MVSREC · Osmania University

Hyderabad · 2023 — 2027 (in progress)

Coursework

  • DSA (Striver A2Z)
  • DBMS
  • Operating Systems
  • Computer Networks
Based
Hyderabad, IN
Graduating
2027
Focus
Full-stack · AI
Status
Open to teams / internships
04

Skills

Grouped by how much I have actually used them, not by how good the list looks.

Comfortable

  • Python
  • Git
  • Linux
  • Docker
  • FastAPI
  • React (components, state, hooks)
  • HTML/CSS

Working knowledge

  • JavaScript
  • TypeScript (basics)
  • SQL (basics)
  • Node.js
  • PostgreSQL
  • Redis
  • pytest
  • Vitest

Used in projects

  • Supabase
  • Traefik
  • WebSockets
  • PyTorch / Real-ESRGAN
  • Gradio
  • Rust (small TUI)
  • npm publish

Familiar

  • Bash
  • Java (basics)
  • C (basics)
  • REST APIs
  • Firebase deploy

Training & certifications

  • AI–ML Virtual Internship

    AICTE · EduSkills (Google for Developers India Edu)

    Jan – Mar 2026PDF
  • Hugging Face Agents Course

    Hugging Face

    2026PDF
  • GitHub Foundations

    DataCamp

    2026PDF
05

Open source & shipping

Small contributions and tools that other people can install and use.

Contribution

Odysseus

Merged PR fixing batch calendar event handling in manage_calendar, part of a public self-hosted AI workspace.

Package

@srinesh/rigit

npm CLI (v0.1.8): one TUI flow for stage, commit and push instead of three separate git commands.

Deployed

Live deployments

Kairo, ReadIssue, Boards, Spentd, pngmodel and this site — all deployed and live.

06

Get in touch

I’m looking for software engineering internships and teams shipping real product. Email is the fastest way to reach me.

rudraksha.srinesh@gmail.com

Replies within 24 hours.