Meet the Maker: Y. Siva Sai Krishna (ysskrishna)

May 20, 2025

Who am I?

I'm Y. Siva Sai Krishna, or ysskrishna online. I'm a full-stack developer based in Bengaluru and an IITM graduate. I've shipped 50+ projects, published 12+ packages on PyPI, and built everything from FastAPI starter templates to AI-powered video dubbing tools. Most of my work lives on GitHub, and all of it is open source.

How it started

I didn't set out to build a large portfolio. The first few projects came from a simple frustration — every time I started something new, I was doing the same setup from scratch. Copy-paste boilerplate, configure Docker, wire up auth, set up CI. Repeat.

So I built starter templates. FastAPI with sync and async patterns. Express + React + TypeScript with SSR and SEO baked in. Next.js + Supabase with auth and RBAC ready to go. Each one solved a problem I kept running into.

Over time, I noticed a few forks and downloads coming in. Not massive numbers, but enough to know that someone out there found the same setup just as tedious as I did — and my templates saved them the trouble.

The PyPI chapter

At some point I noticed I kept writing the same small utility functions across projects. Some were everyday things — normalizing nested dictionaries, extracting social links, converting quote styles. Others came from deeper rabbit holes — detecting file types reliably, stripping byte-order marks from text files, checking if code is running inside a Docker container.

Instead of solving the same thing twice, I started packaging them properly and publishing to PyPI. That turned into 12+ packages — things like text-extensions, binary-extensions, docker-detector, strip-bom, nested-utils, and db-convert. Some are original, some are ports of useful JavaScript/Node.js packages that didn't have a Python equivalent. None of them are glamorous, but they fill gaps I kept running into.

Writing a proper package — with tests, CI, versioning, and documentation — taught me more about software quality than most of my larger projects did.

Where AI comes in

I started building with LLMs and AI APIs when OpenAI's API became widely accessible. My background is backend-heavy, so my angle was always practical: how do you actually build, deploy, and scale an AI-powered application?

Most of the AI projects started as "can I actually build this?" experiments. A RAG-based healthcare assistant to see if I could wire up embeddings with medical data. A video dubbing tool to figure out how Deepgram and ElevenLabs work together for speech-to-speech across languages. A PPT generator because I was tired of making slides manually.

None of them are production-grade products. But each one taught me something concrete — how to manage vector stores, how to handle streaming responses, how to chain LLM calls without the whole thing falling apart. The code is less about the AI and more about the engineering around it.

What I actually use every day

If you look across my 50+ projects, the pattern is clear:

  • Python + FastAPI for backends — type-safe, async-capable, fast enough for production
  • React + Next.js + Tailwind for frontends
  • PostgreSQL when I need relational data, SQLite for lightweight tools, ChromaDB for embeddings
  • Docker for everything — 22 projects have Dockerfiles, and I genuinely enjoy writing multi-stage builds
  • uv for Python project management — replaced Poetry and pip in my workflow
  • GitHub Actions for CI/CD — automated tests, PyPI publishing, deployment pipelines

I've tried a lot of tools over the years, but this stack lets me go from idea to deployed product faster than anything else I've used.

Why I build in the open

Every project I've built is open source. Not out of ideology — but because the best way I learned was by reading other people's code, and I want to return that favor.

When someone forks a template and ships their first API with it, or pulls in one of my PyPI packages to solve a small problem — that's a better outcome than any number of GitHub stars.

What's next

I don't have a grand plan. I'm working on more AI tooling, exploring better patterns for full-stack apps, and trying to get better at writing (hence this blog). The projects will keep coming — I see a problem, I build a solution, and if it's useful, I ship it.

Find me online