Founding Engineer

THE PROBLEM Most products don't fail because of bad engineering. They fail because nobody wanted them. Teams skip research, guess, or follow gut instinct. That's expensive. Replicas is a computational model of human populations: millions of synthetic people with real demographics, real psychographics, and real decision-making patterns. Grounded in census data, behavioral science, live market signals, and real-time consumer sentiment. Always current. No fictional personas. No hallucinated opinions. Statistically validated behavioral simulation at population scale. THE ROLE You'll be one of the first two founding engineers we hire. Half the role is ML: designing, training, and evaluating proprietary models that make our behavioral simulation more accurate, more efficient, and harder to replicate. The other half is systems: building the infrastructure that makes those models usable at production scale for thousands of concurrent users. We're currently experimenting with hybrid LLM and probabilistic simulation systems, and reinforcement-style feedback loops from user interaction data. This isn't a prompt-engineering role dressed up as ML. We are building real model training pipelines, and we need someone who knows what that actually involves. If you believe synthetic populations trained on real human behavior will become a fundamental tool for how products get built and decisions get made, this role will feel obvious. If you're not sure, it probably isn't the right fit. WHAT YOU'LL OWN Model development. Design and train proprietary models for behavioral simulation. Own the training pipeline, evaluation methodology, and research direction alongside the founder and other engineers. Production infrastructure. Own the systems that keep the simulation running fast and reliably as we grow: distributed job processing, real-time updates, horizontal scaling, database performance. You'll make architectural decisions that matter. WHAT YOU'VE DONE BEFORE - Trained and shipped ML models in production, not just API calls or notebooks - Built backend or distributed systems at real scale - Understand human behavior, psychometrics, or decision science deeply enough to model it - Scaled a product and remember what broke when it grew - Seen systems fail under load and understand why - Prefer correctness and signal over superficial outputs - Bonus: worked at a startup that raised strong funding WHAT YOU'LL GET - Compensation: equity-first, with salary introduced post-fundraise - Meaningful founding equity - A technical founder who writes code and has strong opinions - A live product with real users, real data, and real feedback loops from day one - A hard research problem: behavioral fidelity at population scale - Junior engineers and designers under your guidance HOW TO APPLY Before reaching out, spend time with the product. It's free: https://askreplicas.com After trying it, share structured feedback: https://www.askreplicas.com/beta-feedback When you reach out, include: - What you've built that's relevant: ML systems, infrastructure, or anything related to modeling human behavior at scale - Why this problem is interesting to you - Your honest thoughts after using Replicas: what feels strong, what feels missing, and what you would improve We're especially interested in candidates who think critically. Clear conviction matters, but clear critique matters just as much. If you don't resonate with the product direction, this probably isn't the right role. No recruiters. The founder reads everything. Pay: $1.00 - $10.00 per hour Work Location: Remote