Current focus
Data and backend engineer. Co-founded The Boring Rat (acquired). Currently building OpencodeHub, OpenDPP, RAVEN, and JUSTAPI.
Selected work
Live and historical European airspace: ADS-B positions, OpenSky movements, schedules and weather stream through Kafka into a Bronze and Silver lake, a dbt warehouse on DuckDB and a read-only serving copy behind a public dashboard.
A $0 anomaly-detection platform over free satellite and sensor data. Four public sources flow into a versioned Hugging Face lake, dbt and DuckDB build seven gold marts, an IsolationForest scores anomalies, and an edge dashboard serves it from Turso.
GDPR-native streaming and medallion lakehouse for European commerce: real-time fraud scoring, DuckDB and Turso serving, and a six-layer Article 17 erasure cascade with a verifiable audit trail.
OpenCodeHub is a self-hosted, open-source Git platform meant as an alternative to GitHub and GitLab. Git hosting, pull requests, stacked PRs, merge queues, CI/CD, AI-assisted code review, issues, organizations, webhooks, REST and GraphQL APIs, and a CLI all live in one deployable system. It's built as a modular monolith with optional background workers and CI runners, so it runs on one machine and can still scale horizontally. Git works over HTTP Smart Protocol and SSH. Storage is pluggable (local or S3-compatible), several database backends are supported, CI runs in Docker containers, queues use Redis, and code review can talk to multiple AI providers.
A Python web framework with a Rust core: Python runs your handlers while Rust owns networking, TLS, routing, validation and serialization. 766k req/s hello-world, a p99 of 0.48 ms, and a 12 MB footprint.
A modular reasoning architecture that knows when to abstain: six stages with an explicit uncertainty channel, a deterministic DAG router, stateless experts and a verification anchor. GSM8K accuracy 47.2% to 55.6% while confidently wrong answers drop from 19.8% to 2.9%.
Work history
A free, self-hosted alternative to GitHub with SSH access, CI pipelines, and AI-assisted code review, plus a place to discover and share projects. Users run it on their own machines, so their code never touches our servers.
Co-founded The Boring Rat and grew it to 1M+ active users. Built the real-time event ingestion pipelines, the analytics warehouse, and the MLOps workflows, along with the backend systems and infrastructure the product ran on. The company was acquired in November 2025.
Delivered 8+ custom software solutions for international clients, primarily SaaS and e-commerce teams in the US and UK. Designed and maintained PostgreSQL databases, data models, and REST APIs supporting transactional and analytical workloads, including Django APIs serving 10K+ requests a day. Built large-scale web scraping and data ingestion systems, implemented data quality checks and automation, and cut page load times by 40% through React optimization.
Capabilities
40 core · 28 working
Python 
Rust 
Go 
TypeScript 
SQL 
Django 
FastAPI 
Vue.js 
Hono 
Elysia 
Encore.ts 
Node.js 
Apache Airflow 
Apache Spark 
dbt 
Kafka 
AWS Glue 
AWS Redshift 
AWS Athena 
Azure Data Factory 
Google Dataflow 
Google Dataproc 
Google Cloud Composer 
Apache NiFi 
Apache Cassandra 
Apache Druid 
Apache Superset 
Metabase 
MinIO 
Airbyte 
Dagster 
Confluent 
Databricks 
Apache Flink 
Prefect "Deeply curious, very sharp in technical skills, and always eager to dive into complex problems."
"Hardworking and dependable — learns quickly, takes responsibility seriously, and always maintains a positive attitude."
"Swadhin's ability to architect scalable data systems was instrumental in scaling The Boring Rat to 1M+ users. His technical leadership and hands-on approach to infrastructure were key factors in our successful acquisition."
OpenInternet in practice: a private root, a .swadhin TLD, authoritative servers in your closet, and the honest limits of running your own corner of DNS.
20% of my disk free, 2% CPU, 29% RAM, and the machine still hung. Btrfs metadata was 93% full and 125,000 of the files were loose Git objects from an agent snapshot layer.
Efficient pipeline design is mostly a set of decisions about bytes scanned, work repeated, and failures survived. Here is how I reason through them, with the arithmetic shown.
The requirement was 100K rps. The work was arithmetic, a stateless tier, a cache that absorbs most of it, and a long list of things that break before the CPU does.
I built Flare to run mail for my own domain without an IMAP server. Astro on a Worker, Turso for state, Resend or Maileroo for delivery, webhooks for inbound.
Makerspace & visuals
FAQ
Get in touch
Email is the best way to reach me. I'm interested in backend systems, data engineering, AI product work, research tooling, and technical collaborations that are worth the time.
.jpg)