Senior Software Engineer, Mashgin

• 5+ years relevant coding experience, with a track record of shipping production code yourself
• B.S. or higher in Computer Science (or equivalent work experience)
• Strong Python, plus comfort picking up whatever language or tool the problem actually needs
• Hands-on experience with video codecs, encoding, and streaming protocols: video player, media pipeline, or streaming platform work (think the engineering behind a video player at a company like Crunchyroll, Netflix, Twitch, YouTube)
• Experience touching both sides of a pipeline, edge/device code and cloud services, rather than only one layer
• Has built something from zero before: a new service, a new team's first system, a 0-to-1 project with no existing scaffolding
• Comfortable being one of the few people who know how a piece of the system works, at least for a while

Desirable

• We're looking for an engineer to join a new team building features on top of in-store camera feeds (not the kiosk's own sensors, but the store's security/overhead cameras)
• Today that footage mostly just sits on-site, unused
• This team will build the pipeline to get value out of it, whether that means analyzing it where it lives or pulling it into the cloud, across thousands of stores each with its own network constraints and camera hardware
• You'll be an early member of the team
• That means writing the capture/ingestion code, dealing with real codec and streaming decisions (what to encode in, how to transport and play back efficiently over constrained store networks), figuring out what stays on-site versus what moves to the cloud, and getting the first features hooked up
• This has a lot in common with building a video player/streaming stack
• We need someone who's comfortable with some ambiguity: requirements that aren't fully defined yet, tooling you may need to build yourself, and a willingness to just build the thing rather than wait for a full spec
• This is a hands-on coding role first, an infrastructure-design role second
• We maintain a flat, high-impact engineering culture where every engineer who joins our team carries the title of Member of Technical Staff
• Impact: You will own significant responsibility and have total ownership over your work. We stay lean and efficient, so your work directly impacts the product we ship every day
• Simplicity: Complexity is a waste of time here—we take pride in building things that just work
• Hard Work: Forget the glamorous research papers. We prioritize the hard work it takes to achieve real-world clarity and accuracy. You will innovate relentlessly to push the boundaries of what’s possible in visual automation
• Humility: No task is beneath you. Whether it's high-level architecture or grunt work to fix a bug, we check our egos at the door and get real work done
• Ruthless Prioritization: If everything is on fire, nothing is. You will be expected to prioritize ruthlessly, cutting through distractions to focus on the most impactful product improvements
• Customer Innovation: We are customer-centric, not customer-pleasing. You won’t bend to every client demand; you will let innovation drive the product forward to solve the real problems our users face
• Efficiency: Meetings suck. We keep them short and only use them for alignment, giving you your time back to focus on real work
• Writing the actual code, end to end, for a 0-to-1 pipeline: capturing feeds from in-store cameras, encoding, and getting that video where it needs to go, whether that's processed on-site or ingested and stored in the cloud, with no existing system to extend
• Making real codec and streaming-protocol decisions: encoding formats, bitrate/quality tradeoffs, adaptive streaming
• Making fast, pragmatic calls on architecture with incomplete information, and being fine revisiting them as real-world store conditions prove you wrong
• Getting the first feature or two working end-to-end (e.g. a monitoring view, or feeding footage into an early AI feature) to prove the pipeline works
• Working alongside a small, scrappy group across hardware, ML, and backend, figuring out together what "good" looks like, not inheriting a spec
• Dealing with the reality of store hardware: varying camera models, spotty or shared network connections, and physical sites you can't always control
• Doing whatever the moment calls for: debugging a flaky store network connection one day, writing a service to process or ship video the next
• Using AI coding tools to move fast: prototyping quickly, cutting through boilerplate, and iterating on architecture without getting bogged down