About This Opportunity
Asana is building AI Teammates—agents that work like actual users, triaging bugs, responding to requests, drafting briefs, and handling knowledge work across team workflows. Unlike chatbots, Teammates build memory and context across executions. This represents Asana’s shift from tracking work to getting work done.
The team has partnerships with OpenAI and Anthropic and moves at the pace of the industry.
Why I’m Excited
- I’m a Power User - I use agentic tools daily; I understand what works and what breaks
- TypeScript Stack - My primary language, used daily
- QA Background - Evaluation harnesses and reliability are testing problems
- Product Mindset - UX background means I think about users naturally
- Anthropic Partnership - I’m already a Claude Code devotee
Key Qualifications Match
| Requirement | My Experience |
|---|---|
| 4+ years engineering | 7 years across QA, design, and development |
| TypeScript | Daily use, primary language |
| AI/agentic excitement | Daily Claude Code user, prompt orchestration patterns |
| Production systems | Docker, GCP, CI/CD, Playwright automation |
| Product mindset | UX Designer background, cross-functional collaboration |
| Comfort with ambiguity | Adapted workflows as AI capabilities evolved |
Unique Angle
Most engineers applying haven’t spent hundreds of hours as end users of agentic systems. I have. I understand:
- When agents need guardrails vs autonomy
- How memory and context affect user trust
- What makes multi-step planning feel reliable vs brittle
- Where human-in-the-loop checkpoints matter most
This user intuition, combined with my QA foundation for building evaluation systems, is directly applicable to building AI Teammates.
What I’d Build
Based on the job description, I’m particularly excited about:
- Evaluation harnesses - My QA background makes this natural
- Agentic workflows in production - I use these daily, now I’d build them
- Memory and orchestration - Core to what makes agents useful vs frustrating
- A/B tests and experiments - Product-minded approach to measuring impact