Website:
workelate.com
Job details:
Our mission
The phrase "AI-native" is now a marketing checkbox on most products. We mean something specific: documents that retrieve their own context, sheets that explain their own anomalies, presentations that update themselves when the underlying data changes, and agents that act across the suite with full memory of the org.
This role owns the AI substrate that makes that possible: orchestration, retrieval, memory, evaluations, and the agent runtime.
What you'll own
- The agent and tool-use runtime across all surfaces (docs, sheets, mail, chat, tasks, journeys).
- Retrieval architecture - how WorkElate indexes, embeds, retrieves, and ranks across an org's entire knowledge.
- Model orchestration - when to use which model, latency budgets, streaming, fallbacks, cost controls.
- Memory and personalization - short-term context and long-term per-user / per-org memory.
- Evals - the test suite that lets us ship AI features without breaking trust.
Who we're looking for
- 4+ years shipping production AI/ML systems, ideally including LLM-based products in the last 18 months.
- Hands-on experience with retrieval (vector + keyword + structured), agent frameworks, and streaming inference.
- You have run real evals - not vibes. You can explain how you measured a model change before shipping it.
- Strong distributed systems background. Latency, queuing, backpressure are second nature.
- You ship. Your bias is to put working agents in users' hands and iterate.
Bonus signals
- Worked on AI products at Anthropic, OpenAI, DeepMind, Cohere, Mistral, or top AI startups (Cursor, Replit, Perplexity, Glean, Harvey, Hebbia, etc.).
- Open-source work on retrieval, agents, or eval frameworks.
- You have written publicly about why most "AI in productivity" features feel hollow and what would actually make them feel inevitable.
First 90 days
- Day 30 - First version of the WorkElate AI orchestration running across apps, with a clean interface.
- Day 60 - Org-level retrieval working across; eval harness in CI.
- Day 90 - At least three flagship AI features shipped to design partners with measurable quality wins over baseline.
Compensation
Founding equity 1.5–3.0%. Cash compensation was discussed in the first call and benchmarked to your geography and current package. We are not the highest-cash offer you will get ( 40-50K). We intend to be the highest-equity, highest-leverage one.
How to apply
1. Fill the form with the intent and various work exposures you have
2. Send a 60–90 second Loom answering: "What's the one decision in editor architecture that most teams get wrong, and why?"
3. Include a link to code, a system you've designed, or a write-up that shows how you think.
Email: teams@workelate.com— subject line "Founding Engineer — AI".
Click on Apply to know more.