Data Scientist, KOHO Financial

• You have strong fundamentals
• You're fluent in Python, comfortable with the full ML stack (data wrangling, feature engineering, model training, evaluation, deployment), and you know when to use a simple model and when complexity is warranted
• You think in systems
• You're not satisfied shipping a model that works in a notebook
• You think about data pipelines, drift, monitoring, retraining triggers, and what happens six months after launch
• You build things that last
• You own outcomes, not outputs
• You don't consider the job done when the model is deployed
• You track what's happening in production, investigate regressions, and hold yourself accountable to the business metric, not just the eval metric
• You're collaborative without being dependent
• You can drive a project independently and also know when to pull in the right person
• You communicate clearly in writing, contribute to design discussions, and give honest feedback
• You care about craft
• Clean code, reproducible experiments, documented decisions
• You understand that how you build matters because other people have to work with what you leave behind
• Experience range: We're flexible. If you're 3 years in and have owned real production models, we want to hear from you. If you're 8 years in and want to work on harder infrastructure problems, same

Desirable

• We're building a model acceleration platform: the reusable infrastructure, tooling, and practices that let KOHO ship high-quality AI and ML faster, with more confidence, at every level of the stack. That means owning models end-to-end, building things that compound over time, and being the team that other engineers and product teams lean on when the problem is genuinely hard
• I'm looking for a Data Scientist who wants to work at that intersection: deep technical craft, full lifecycle ownership, and real product impact. Someone who thinks in systems, not just models. Someone who cares about how the work is done, not just whether the notebook ran clean
• Design, build, and ship ML models that solve real product problems across KOHO's core surfaces: spending, savings and lending
• Own your models in production: monitor them, debug them, iterate on them as the world changes
• Frame problems rigorously before reaching for a model: ask whether ML is the right tool and what success actually looks like
• Build and improve shared ML infrastructure: feature stores, experiment tracking, model registries, evaluation frameworks, monitoring pipelines
• Contribute to making the platform genuinely self-serve for other teams, reducing friction in the model-to-production path
• Help establish reusable patterns that compound across projects, not just solve today's problem
• Apply responsible AI principles throughout development: fairness evaluation, model explainability, documentation, and audit-readiness
• Work closely with ML engineers, data engineers, product managers, and analytics engineers to translate ambiguous business problems into scoped, solvable technical work
• Participate in design reviews, contribute to architectural decisions, and hold a high bar for how the team builds