Lead Software Engineer & Solutions Architect based in Auckland, NZ.
I build distributed systems, data pipelines, and AI-native infrastructure and I like doing it in the open.
- VecLite: client-side vector search that scales. Rust/WASM powered, 100k vectors in 400ms, entirely in the browser. No server. No API keys. No data leaves the device.
- IoT telemetry at scale: Go + Kafka + TimescaleDB pipeline handling 16K msg/s
- LLM-native tooling: agentic workflows, RAG pipelines, MCP servers
- DevSecOps patterns: security-first architecture from day one, not bolted on at the end
Client-side vector search that scales. Rust/WASM powered, works where pure JS breaks down.
import { VecLite } from 'veclite'
await VecLite.init()
const db = new VecLite({ dimensions: 1536 })
db.upsert([{ id: 'doc1', vector: embedding, metadata: { category: 'science' } }])
const results = db.search({ vector: queryEmbedding, topK: 5 })
// 100k vectors. 400ms. Zero server.
→ Check it out · PRs and issues welcome · v0.4 coming soon with veclite/rag
- Contributing to OSS projects in Go, TypeScript, or Rust: especially distributed systems, data infra, or AI tooling
- Collaborators on VecLite and other projects: PRs, issues, and ideas are genuinely welcome
- Mentoring engineers navigating the jump from mid to senior, or into architecture
- Building together: if you have an interesting problem, let's talk
If you want the full picture, past work, case studies, and what I'm thinking about -
→ scalewithaakash.com
Or just say hi: tmaakash143@gmail.com