Skip to content
View suryansh98's full-sized avatar
🎃
Chipi chipi
🎃
Chipi chipi
  • India
  • 19:36 (UTC +05:30)

Highlights

  • Pro

Organizations

@Neko-Nik-Org

Block or report suryansh98

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
suryansh98/README.md
Suryansh Chaturvedi

LinkedIn Email India


Hello

I'm Suryansh. I started writing code at 13 and never really found a reason to stop.

What I like is the layer underneath the framework — storage engines, write-ahead logs, query planners. The stuff you only understand properly by building one badly first, which is roughly how the database below came about.

Off the keyboard I'm usually watching anime, or losing an evening to someone's blog post about how their database survives a power cut. Occasionally both at once.


Featured work

A SQL database you can see inside

An embedded database engine written from scratch in Go — with zero third-party dependencies. Standard library only.

Retrieval-augmented QA over PubMed

A four-person MSc NLP project at Universität Heidelberg. I built the LLM answering layer.

  • Paged storage, B+tree indexes, LRU buffer pool
  • Write-ahead log, SQLite-style crash recovery
  • Per-frame CRC and torn-write detection
  • Recovery proven by a fault-injecting filesystem
  • Hand-written SQL parser and query planner
  • Live event-stream visualiser (xray)
  • PubMed abstracts scraped, cleaned and chunked
  • Chunk embeddings indexed in a Pinecone vector DB
  • Answering layer on Llama-2-13b-chat
  • Evaluated against a generated question set
  • Retrieval and generation split into own stages
  • Team project, written up as an MSc report
~7k LOC  ·  76% coverage  ·  0 dependencies
Go, React
Python  ·  Llama 2  ·  RAG  ·  Pinecone
Python

Tools of the trade

Languages TypeScript, JavaScript, Go, Python, PowerShell, Bash
Frontend React, React Native, HTML, CSS, Figma
Backend & Data Node.js, Express, GraphQL, PostgreSQL, MySQL, MongoDB
Cloud & DevOps AWS, Azure, Docker, GitHub Actions, Linux, Git
Observability & Testing Grafana, Loki, Sentry, Jest, Playwright, Postman

REST · System Design · CI/CD · Agile/Scrum · SAST · Playwright · React Native · Loki · LLM · RAG


By the numbers

Profile summary
Repos per language Most commits by language
Contribution streak

Contribution graph, eaten by a snake

Snake eating my contribution graph


Get in touch

Always happy to talk about storage internals, payments plumbing, or anything with a hard systems problem underneath it.

suryansh.official100@gmail.com  ·  LinkedIn



Pinned Loading

  1. agent-pipeline agent-pipeline Public

    Local web app and CLI for repo-aware AI coding workflows, running against your own repos and git checkout.

    TypeScript

  2. Alternative-Youtube Alternative-Youtube Public

    A minimal, distraction-free YouTube client with Google OAuth, subscriptions and a clean feed. React, TypeScript, Express.

    TypeScript

  3. biomedical-rag-qa biomedical-rag-qa Public

    Retrieval-augmented QA over PubMed abstracts. MSc NLP team project, Universitat Heidelberg WS2023.

    Jupyter Notebook

  4. Bomberman-MLE-2024-Mlbot Bomberman-MLE-2024-Mlbot Public

    Q-learning Bomberman agent with hand-engineered features. MSc Machine Learning Essentials project, Universitat Heidelberg.

    Python

  5. glassdb glassdb Public

    A small SQL database you can see inside. Embedded engine in Go with B+tree storage, WAL crash recovery, and a live X-ray visualizer.

    Go

  6. youtube_summarize youtube_summarize Public

    Turn any YouTube video into an AI summary: bullet points, detailed notes or key takeaways, in 12+ languages.

    TypeScript