Software engineer & founder. I build distributed systems for fintech, Web3, and AI — and I care about the parts you don't see: reliability, observability, and year two of maintenance.
LinkedIn · X / Twitter · Medium · Email
Software Engineer · Kryptos — Göteborg, Sweden Enterprise crypto accounting and tax. I work on high-volume, multi-chain transaction ingestion and the compliance engine that turns messy on-chain data into audit-ready numbers.
Co-founder & CTO · Aura AI AI-powered makeup consultant and virtual stylist. I own the architecture end to end — the computer-vision pipeline for face and skin-tone analysis, the recommendation engine, and a privacy-first design where images are processed in memory and never written to disk.
Systems that ran in production with real users and real money — not technologies I've merely touched:
- Identity at scale — OIDC/OAuth2 platforms: token lifecycles, session security, multi-tenant SSO.
- Event-driven microservices — ordered, idempotent, replayable pipelines that survive partial failure.
- Crypto compliance tooling — deterministic cost-basis and tax computation over inconsistent multi-chain data.
- AI-powered automation — NLP and LLM pipelines with evals and fallbacks, built to ship rather than demo.
- Solana programs — Anchor/Rust on mainnet, where a bug is a financial event.
- Boring technology, chosen deliberately. Novelty is a cost; I spend it only where it buys leverage.
- Design for the failure case first. The happy path is easy — retries, idempotency, and backpressure are the actual design.
- Observability is a feature. If you can't see it in a dashboard at 3 a.m., it isn't done.
- Code is read far more than it's written. I optimize for the next engineer, including future me.
| Area | Daily drivers | Also fluent |
|---|---|---|
| Languages | TypeScript, Python, Go | Rust, C++, SQL, Bash |
| Backend | Node.js, FastAPI, NestJS | Gin, tRPC, GraphQL, WebSockets |
| Frontend | React, Next.js | Vue 3, React Native |
| Data | PostgreSQL, Redis | MongoDB, TimescaleDB |
| Infra | AWS, Docker, Kubernetes, Terraform | GCP, GitHub Actions |
| AI / ML | OpenAI & Anthropic APIs, LangChain, PyTorch | scikit-learn, OpenCV |
| Web3 | Solana (Anchor / Rust), Web3.js | Ethereum |
Background: 5+ years in production software · CGPA 9.3/10 · Amazon ML Summer School · Certified Network Defender · 4★ CodeChef
A few things I've built in public — small, focused tools that solve a real problem:
- SolSyncer · TypeScript — A bridge between on-chain Solana events and off-chain job execution via BullMQ, for DeFi protocols that need reliable real-time data processing.
- dreamingai-mcp-server · TypeScript — MCP server exposing an enterprise AI assistant's conversations, agents, and models over the Model Context Protocol.
- LiquiDash · JavaScript — Smart contracts, TypeScript SDK, and dashboard for a Liquity-style governance-free borrowing protocol (interest-free ETH-backed loans).
- npm-uninstall-unused · JavaScript — CLI to detect and remove unused npm dependencies across single projects and monorepos.
- cursor-storage-reset · Python — Cross-platform CLI to safely reset Cursor telemetry IDs, with atomic writes.
I write about systems, fintech infrastructure, and lessons from production on Medium.