Come build the next Web Revolution
We're making AI chatbots accessible to every business on the planet , trained on their content, running on their users' devices, at a fraction of the cost. Small team. Big ambition. No bureaucracy.
Open Roles
Three roles. Three ways to leave a dent.
Product Engineer
Part Designer. Part Frontend. Part PM. Full Menace.
You own the surface area , from user research to pixel-perfect UI to shipping features that move metrics. You'll design and build the dashboard, the embeddable chat widget, and the showcase experience. You care about craft, you ship fast, and you have opinions about what to build next.
What you'll do
- Design and build user-facing features end-to-end
- Own the frontend architecture (Next.js, React, Tailwind)
- Run lightweight user research to inform product direction
- Ship fast, iterate faster , we don't do roadmap theater
ML Engineer
From raw crawls to trained models. The full pipeline.
You turn scraped web content into production-grade chatbot models. Data gathering, data cleaning, QA pair generation, post-training optimization, and prompt engineering , you own the intelligence layer. You'll work with small models that need to punch way above their weight.
What you'll do
- Build and optimize the training data pipeline
- Fine-tune small language models for on-device deployment
- Design prompt templates and evaluation frameworks
- Push the boundary of what tiny models can do
Platform Engineer
Rust backends. Kubernetes clusters. The whole stack.
You build the backbone , the Rust API, the crawler infrastructure, the training orchestration, the deployment pipeline. Headless browsers, Redis queues, Kubernetes, Terraform , you keep the engine running and make it scale. Systems thinking is your default mode.
What you'll do
- Extend and optimize the Rust/Axum backend
- Scale the crawling and training infrastructure
- Manage Kubernetes, Terraform, and CI/CD pipelines
- Design systems that are simple, fast, and reliable
Why Kanha?
Ship, don't slide
We build and ship. No decks, no committees, no approval chains.
Small team, big leverage
Every person here moves the needle. You won't be a cog.
Hard problems
On-device ML, browser-based inference, training pipelines , real engineering.
Remote & async
Work from anywhere. We care about output, not hours logged.
Don't see your role?
If you're exceptional at what you do and want to work on the frontier of on-device AI, reach out anyway.
Say hello