AI Engineer - BigUp

About Us

BigUp is building a new category of entertainment: Predictainment . Attention has shifted from TV networks to creators, and prediction markets have transformed how audiences engage with live events. We're building a platform that lets any creator launch their own TV network — think Twitch meets Kalshi.

Creators host livestreams supported by proprietary AI agents that handle production, clips, stats, and graphics, while fans watch and predict alongside their communities in a multiplayer experience. We're a small, scrappy team of builders from Stanford, MIT, Circle, and Microsoft, backed by top-tier VCs and industry leaders from OpenAI, DraftKings, Whatnot, and Amazon Studios.

We move fast, ship constantly, and figure things out as we go — there's no playbook for what we're building. We just came off a successful pilot and we're now racing into new markets — starting with NFL and college football, then non-sports verticals this fall. We need an AI Engineer who wants to build the agent infrastructure from scratch, not tune models on top of someone else's system.

Job Description BigUp is looking for an engineer to help build a category of live streaming that doesn't exist yet — one where AI agents work alongside creators in real time to produce, enrich, and gamify every broadcast.

You'll build the agent harness that powers this experience, orchestrating AI agents that handle production, clips, stats, and graphics live, and the data ingestion pipelines that feed them accurate, low-latency information from live sporting events. There's no playbook for this.

You'll be architecting systems that let AI keep pace with the unpredictability of live events, at the speed and scale a real audience demands — often before anyone's fully speced out how they should work.

You'll work directly with our CTO, not through five layers of process, and the systems you build will directly shape how creators broadcast and how fans predict alongside them as we expand from live sports into an entirely new form of interactive entertainment. If you want a role with a clear spec and an established system to optimize, this isn't it.

If you want to build the agent infrastructure that doesn't exist anywhere else yet, keep reading.

Responsibilities

Design, build, and own the agent harness that powers our production agents — end to end, from architecture to production Build and scale real-time data ingestion pipelines that pull live data and route it to agents with low latency Implement eval frameworks to monitor agent output quality and debug performance issues, often in real time during live broadcasts Work directly with product design to turn creator and viewer needs into agents that just work — no lengthy spec process in between Integrate live video playback and real-time data feeds (prediction markets, live stats, overlays) Work directly with backend engineers to design APIs as we go, not after the fact Qualifications 3-4+ years shipping real production systems — degrees are nice, shipped products are better A track record of building things on your own time — side projects, indie tools, open source, hackathon wins, whatever it is.

We care much more about what you've built unprompted than which specific frameworks you've used Comfortable picking up new tools fast — TypeScript/React experience helps, but we're more interested in how quickly you learn than a checklist of technologies You've built agent harnesses in production before (tool calling, evals, etc.) or you've clearly done the equivalent on your own Comfortable with streaming/event-driven architectures (Kafka, Kinesis, or similar) Genuinely energized by ambiguity and fast pivots, not just tolerant of them You can communicate clearly and move quickly without a lot of process around you You're an actual sports fan or prediction market user — this should feel like building for yourself, not building for a market you read about The Perks Competitive salary + meaningful equity — real ownership in a company at an inflection point A small, high-caliber team from places like Stanford, MIT, Circle, Microsoft, OpenAI, DraftKings, Whatnot, and Amazon Studios The chance to build a category-defining product from near-zero, not maintain someone else's Health insurance Generous PTO How to Apply DM us a l ink us to something impressive you've built (GitHub, App Store link, portfolio, etc.) — we care way more about this than your resume.

Share an interesting fact about your favorite creator and why you'd like to see them on BigUp.