Full-time. San Francisco, in person. Base salary: $120,000 to $240,000 per year. Equity: approximately 0.5% to 1.5%, depending on experience and scope. ## Build agents that carry work through to a useful result Screenpipe helps companies find work worth automating. It captures screen and audio activity, makes computer history searchable, and gives AI agents context from how work happens. Capture history stays on the device by default; optional cloud AI, sync, and enterprise storage have separate data paths. We want agents to do substantial work across engineering, support, research, sales, and operations. People set direction and make decisions that need human judgment. You'll work directly with the founder on the product and the agent infrastructure we use ourselves, then turn what works into capabilities customers can rely on. ## What you'll own - Build persistent memory, context retrieval, tool integrations, MCP servers, and orchestration for work that continues across sessions. - Make workflows recover from interruptions, retry safely, and verify results without duplicating external actions. Define permissions and when a person needs to take over. - Build traces, reproducible failure cases, and focused regression evaluations so we can understand failures and measure improvements. - Improve Screenpipe's desktop capture, local storage, search, APIs, and interfaces. Investigate crashes, audio failures, memory use, and CPU usage on real machines. - Talk to users, choose a useful scope, ship it, and follow through on whether it helped. Keep the code understandable as the product grows. ## What we're looking for You've personally shipped software and can explain a difficult bug, the tradeoffs you made, and the outcome. You can debug unfamiliar code and take responsibility for changes produced by coding agents. You use AI heavily in your own work and experiment with models, tools, and automations. You understand the relationship between token usage, cost, latency, and useful output. We assess the systems and results behind that usage; there is no fixed token-spend, degree, or years-of-experience threshold. Our stack includes Rust, TypeScript, React, and Tauri. Bring depth in part of it and the ability to learn the rest. The work spans product interfaces and backend or systems engineering. You'll need initiative, direct communication, and a willingness to change your mind when evidence warrants it. This role is in person in San Francisco. ## How we work Each role has a scorecard with clear objectives and measurable targets, so you know what you're responsible for and what success looks like. We review it together every two weeks, look at the results and what's getting in the way, and agree on the next priorities. You'll own how you reach those goals and speak up when the evidence calls for a different approach. We work long days, and weekend work is part of our pace. We want you to understand that expectation before joining. We also make time to celebrate wins together and recognize the work that got us there. ## Benefits - Health and recovery: Personal health coach or trainer, gym membership, sauna and cold plunge access, and unlimited massages. - Health insurance: Top-tier health insurance. - Flexibility: San Francisco is our home base, with flexibility to work remotely from time to time, including trips abroad such as Thailand. - Learning: Unlimited books and audiobooks, on us. - AI tools: $20,000 per month in AI tokens to build, experiment, and do your best work. - Personal growth: Career and life coaching. - Team travel: Team trips and work retreats, including places like Hawaii. ## Apply Keep answers concise. Links and bullets are welcome. Include your current location and when you could start working in person in San Francisco. Use public or anonymized examples, and remove credentials, private prompts, and confidential customer or employer information. 1. Link two or three things you shipped. Explain your contribution, a difficult bug or tradeoff, how you investigated it, and an outcome you verified. 2. Show a workflow that runs beyond a single prompt. Walk through its trigger, tools, saved state, output checks, recovery, and human handoff. Include your agent setup and approximate daily or weekly token usage or spend, distinguishing personal usage from team usage. Explain what that usage accomplishes. 3. What are you learning now? Share a book or idea that changed how you work, your two or three favorite books, and roughly how many you read or listened to in the past year. Include an example of feedback that changed your mind. 4. What would you automate first inside Screenpipe, and how would you know it was working? Explore the product at https://screenpipe.com and the codebase at https://github.com/screenpipe/screenpipe.