Hi, I’m Yogesh, full-stack engineer.

I design how the pieces connect.

I build products end to end, and I care most about the parts in between: the pipelines, queues, tunnels and protocols that decide whether a product actually works. Lately I go deep on a small set of tools instead of skimming ten stacks.

SDE intern at SystemStar Infotech · B.Tech CSE at Heritage Institute of Technology, ’28

$ curl -w "@timing" -so /dev/null yo9esh.tech
dns lookup
…
tcp connect
…
tls handshake
…
first byte
…
download
…
page ready
…
Real numbers from your browser’s Navigation Timing API. A 0 means the connection was reused.

Selected systems

05 case studies
01

Chatterbox TTS pipeline · 2026

Fitting a 48 GB voice model into 5 GB

Turning thousands of real-estate listings into AI voiceovers. The speech model’s memory grew with input length, so long listings crashed or hallucinated, and it needed 48 GB-class GPUs to survive.

  • 16–80 → 5 GBVRAM, any input length
  • −87%infra cost
  • 2,655+listings, zero OOM
Constraint
S3Gen held the whole sequence in memory: 16 GB at baseline, 48–80 GB on long inputs. Position overflows at token boundaries also produced garbled “hallucinated” audio.
Decision
Rewrote S3Gen to run in bounded-memory chunks and patched the position and token-boundary handling, so memory stays flat regardless of length.
Tradeoff
Chunking adds stitching logic at the boundaries, but it gives memory a hard ceiling, and that ceiling is what moved inference from 48 GB accelerators to an RTX 3070.
PythonPyTorchCUDARunPodMySQLBunnyCDN
02

SystemStar Infotech · 2025–26

A 360° player that starts fast

4K 360° video is heavy, and every second before the first frame loses a viewer. I built the player and the pipeline behind it.

  • 4K360° source footage
  • 2 sHLS segments (vs 4 s)
  • ABRadaptive bitrate ladder
Constraint
Generic CDN presets use ~4 s segments, so the player waits for a big first chunk of an already huge 4K stream before showing anything.
Decision
Custom preprocessing that segments streams into 2 s chunks across an adaptive bitrate ladder, played on a Three.js sphere with modular camera controls.
Tradeoff
Shorter segments mean more files and more playlist requests. In return, the first frame arrives sooner and quality switches react faster.
ReactThree.jsFFmpegHLSCDN
03

CoralTeams · Coral hackathon · 2026

LAAF: infra from a potato laptop

A self-hosted workspace search across GitHub, Jira, Slack and Docs. It needed a 1 GB+ Rust engine plus a 3 GB local LLM, and my laptop couldn’t run either.

  • $0per month
  • 3free clouds, 2 tunnels
  • 4 GB+stack, off-laptop
Constraint
No high-spec laptop, and cloud free tiers kept rejecting card verification. The stack couldn’t even be tested locally.
Decision
What I call CNG: SSH into GitHub Codespaces for the backend, run the LLM on Google Colab’s GPU, and expose both through ngrok. The laptop only renders the UI.
Tradeoff
Colab sessions expire and tunnel URLs rotate, so it’s a dev and test rig, not production. The same containers move to one persistent server later.
04

1st prize · Ideatex, GDG on Campus HIT · 2025

Cursor for n8n

Describe an automation in plain English, get a working n8n workflow. Built in five days with Soumyadeep Bhowmik.

  • Firstof 400+ developers
  • 5 daysidea to pitch
  • RAGover real node specs
Constraint
LLMs invent n8n node names and parameters, so the generated JSON looks right and fails on import.
Decision
An MCP server that retrieves real node definitions from n8n’s own SQLite knowledge base and grounds the model before it writes a single field.
Tradeoff
Grounding costs a retrieval step on every request. In exchange, workflows import and run on the first try.
TypeScriptNode.jsMCPRAGSQLite
05

Open source · GSSoC ’25 Project Admin

AlchEmaiLyst, and the community around it

A three-agent email assistant that summarises, composes and sends. Then I opened it up through GirlScript Summer of Code and ran it as a project.

  • 42★GitHub stars
  • 30+contributors onboarded
  • #1Alchemyst leaderboard
Constraint
Reading someone’s inbox means OAuth tokens and LLM API keys. Neither can ever reach the browser.
Decision
A thin Express proxy owns Google OAuth 2.0 and every LLM call. The UI only ever talks to the proxy.
Tradeoff
One more service to run, in exchange for a single place to secure, rate-limit and swap model providers.

Also built

  • Fullstack Odyssey Interactive simulators for the event loop, SQL joins, JWT rotation, Redis queues and Git’s DAG.
  • Lost & Found Campus lost-and-found with Gemini auto-tagging and a scored matching engine. Built in 2 hours; won.
  • Receipts.sh Telegram bot that logs promises in group chats, nudges before deadlines and deletes leaked API keys.
  • Moot Spiral Studio Photos + music → 60 FPS Fibonacci-spiral video, rendered entirely in the browser.
  • 3dWebImg Any 2D image into a 360° panorama, fully client-side.
  • ClimaVisor Climate forecasts and history on an interactive 3D globe.

Systems notebook

notes from the build

On the podium, in the room

20+ hackathons
SIH 2024 Grand Finale: KrishiSahayak · Ministry of Fisheries, Animal Husbandry & Dairying
SIH 2024 Grand FinaleKrishiSahayak · Ministry of Fisheries, Animal Husbandry & Dairying
Ideatex, 1st prize: GDG on Campus HIT · Session Zero 2025
Ideatex, 1st prizeGDG on Campus HIT · Session Zero 2025
e-Summit ’25: two podiums: EDC, Heritage Institute of Technology
e-Summit ’25: two podiumsEDC, Heritage Institute of Technology
TCS hackathon win: A two-hour build, after 8+ tries with the same partner
TCS hackathon winA two-hour build, after 8+ tries with the same partner
Recognition
Jul 2026 Winner · TCS hackathon · Lost & Found, end to end in 2 hours
2025 1st runner-up ×2 · e-Summit ’25, EDC HITK · Ideathon (Optiq) + Marketing Gimmick
Jul 2025 1st prize · Ideatex, GDG on Campus HIT · Cursor for n8n · 400+ developers
Feb 2025 2nd runner-up · ACM Brainstorm quiz
Dec 2024 Grand Finalist · Smart India Hackathon · Top 0.2% of 50k+ teams · team KrishiSahayak
Sep 2024 3rd of 200 · HackHeritage · SIH college prelims

Few tools, all the way down

depth over breadth
ssh
Remote dev in Codespaces over gh codespace ssh; port forwarding to reach services on remote boxes.
ngrok / tunnels
Joined three free clouds into one dev stack for LAAF.
curl & HTTP
Debugging APIs from the terminal; the timing card at the top of this page is the same idea.
FFmpeg + HLS
The 4K → adaptive-stream pipeline at SystemStar.
TypeScript · Node
MCP servers, Express proxies, OAuth flows.
Python · PyTorch · CUDA
Rewrote model inference to fit an 8 GB GPU.
Postgres · MySQL · Redis
Data and queues behind CoralTeams and the TTS pipeline.
React · Three.js
The 360° player; most of the interfaces above.

Right now: running my own VPS end to end (systemd, Caddy, Docker), Postgres internals, queues with retries and idempotency, and reading Designing Data-Intensive Applications.

Yogesh in a yellow kurta at an old Kolkata gate, smiling

About

From Gaya to Kolkata, one shipped thing at a time.

I’m a Computer Science undergrad at Heritage Institute of Technology, Kolkata, and a software engineering intern at SystemStar Infotech. I like the unglamorous middle of a product: how 4K video starts playing on a phone without the wait, how an LLM gets grounded in real data, how a stack runs when the hardware says it can’t.

How I work: I show work early. I got my internship by building a mock before the interview, not after. Here’s that story.

I’ve shipped in hackathon rooms more than 20 times, from a two-hour build to the SIH Grand Finale, and I run open-source projects too. Off the keyboard, I sing and play volleyball.

Contact

Have a system that needs its pieces connected?

fbyogesh111@gmail.com