Harness Engineer

The Founding Engineer owns and builds the core product while Joseph pushes the leading/experimental edge — the two are a redundancy pair who cover for each other. The absolute must is harness engineering: this is a "content engineer / harness engineer" role, not a traditional software-dev one. The mental shift Khailee wants: a normal engineer sees a bug and goes fixes it; this person designs a system that improves itself — e.g., every user feedback click logs a screenshot + full error context, and a standing agent takes the nightly error logs, fixes things itself, and reports back what it improved for review. They must be deeply plugged into the frontier of models (niche/Hugging-Face models good at X or Y) and know how to integrate them so it feels seamless to the user, and eventually liaise closely with the ML hire to train small on-device models for parts of the creative pipeline (e.g., a brief-taking agent, a QA agent).

Profile shape: a T-shaped, well-rounded generalist (solid front and back end, not hyper-specialized) whose deep spike is harness engineering — a "natural-language genius" who holds the whole thing in their head, not a tech leader/manager. The next six months are about four levers on the core product: bring down the time a video takes (currently ~15 min to 13 hrs; use faster models, show quick in-between drafts for dopamine), increase the dopamine/magic (integrate the best niche models), bring down cost per video (voice-clone+lip-sync renders at ~1/10 cost, cache repeated web searches), and build the self-improving/recursive harness for stability. What matters most to Khailee: what they've built — someone who's "gone through the fire" of something similar (core engineer on a comparable product, core contributor to a relevant open-source project).

What You'll Own

• The core product end to end, paired with Joseph (redundancy — cover for each other)
• Harness engineering — designing self-improving systems (nightly error-log-driven agent that fixes and reports back), not just fixing bugs
• Integrating frontier/niche models (Hugging-Face, model-per-task) so the pipeline feels seamless and fast
• The four six-month levers: cut time-to-video, raise dopamine/magic, cut cost per video, build recursive stability
• Liaising with the ML hire to train small on-device models (brief-taking agent, QA agent, etc.)

Requirements

• Harness engineering — a 100% must (the deep spike; the whole role hinges on it)
• T-shaped, well-rounded generalist — solid across back end and front end, not hyper-specialized
• Deeply plugged into the frontier: niche/open-source models, model integration, agentic pipelines
• A "brainiac" IC who holds the system in their head — not a tech-lead/manager profile
• Has built something similar and "gone through the fire" of it (core engineer / core OSS contributor on a comparable product) — company/school irrelevant

Execution and Ownership

• Thinks in self-improving systems, not one-off fixes
• High agency, intense, all-in; genuinely interested in the video/creation problem (intrinsic motivation is screened for)
• Low ego ("non-positional" — builds shared understanding, doesn't defend turf)
• New-school engineer who harnesses AI to get the job done, not an old-school "do it all by hand" one

Background

• Slope/craft over seniority — explicitly not the "18 years, led 90 people" old-school/scale profile
• Newer-school audience welcomed; what they've built >> pedigree
• Renaissance/creative differentiator (nice-to-have, not required): also expresses a creative side — plays an instrument, makes content, real artistic intuition — a strong tiebreaker among equals (but they have enough creatives; competence first)

Who Will Thrive Here

• A harness-engineering specialist who designs self-improving systems and lives at the model frontier
• A T-shaped generalist IC who holds the whole core product in their head — not a manager
• Someone who's built something similar and went through the fire of it
• All-in, intense, low-ego, in a concentrated-intent life phase; weekends-on
• Bonus: a genuine creative (instrument, content) who intuits the art side as well as the engineering