Languages
TypeScript everywhere it fits. Python for ML / data. Go for high-concurrency backends. Rust when it earns the complexity.
The boring tech we love.
Our default stack is the one that gets out of the way. We pick newer tools when they earn it — never just because they're new.
Default stack
Frontend
React, Next.js, Remix, Tailwind. We don't ship class-based components in 2026.
Backend
NestJS, Express, FastAPI, Go (chi/echo). Picked per workload, not per fashion.
Data
Postgres for OLTP, ClickHouse for analytical, Redis for cache + queues. We avoid Mongo unless the schema really is document-shaped.
Infra
AWS first, GCP when client-mandated. Terraform, GitHub Actions, Vercel for fronts. Kubernetes only when the team can run it.
AI / LLM
Claude for reasoning, OpenAI for cheap-and-fast, fine-tuned open models when latency / privacy matter. We pick the boring tool.
Tools we use day-to-day
- TypeScript
- React
- Next.js
- Postgres
- AWS
- Terraform
- Linear
- Figma
- Sentry
- Datadog
- GitHub
- Vercel