Together AI product team opens up its real repo: full AI coding workflow revealed
aakashgupta · x · 2026-09-15
In a nearly hour-long interview, Together AI's product team—the company just raised $800M at an $8.3B valuation—shows the shared repo they actually run on.
Key takeaways:
- Moving from individual outputs to a shared product repo: who owns context files, team skills vs personal skills
- Choosing a harness and model, and building your own orchestrator
- A PRD skill that interviews you, compressing half a day of research into 5 minutes, plus a customer insights MCP
- Agent evals on shipped features, design validated every few hours
- Core insight: context is a hierarchy, not a pool; what stays human in the loop
- What the whole setup actually cost them
More from coding & agent
- The detective novel thought experiment: why hidden states beat tokens for model handoff — CShorten30 · 2026-09-15
- oh-my-pi open-sources agent context compaction design docs, now at 31k GitHub stars — nirmal_dist · 2026-09-15
- RAFT v3.1 open-sources retrieval-augmented fine-tuning to clone human personas — jessi_cata · 2026-09-15
- 18 sub-agents cut to 1: Polylane argues sub-agents are just wrong — zeeg · 2026-09-15
- deepseek-harness: an open-source plugin-based coding agent framework, unverified — goyalshaliniuk · 2026-09-15
- DeepSeek's full-stack code output lags Claude Code and Cursor — devs trade prompting tips — AdEnvironmental4143 · 2026-09-15