Oneapp The Applied Scientist, Foundation Models at Oneapp designs, trains, and deploys foundation models for fintech applications. This role fine-tunes models, mitigates risks, and optimizes search and retrieval systems.
Responsibilities
As an Applied Scientist, Foundation Models at OnePay, you'll be a founding member of our foundation model effort, building it from the ground up. You will:
Design, train, and evaluate OnePay’s foundation models, including models built from scratch on our proprietary data and adaptations of existing open or commercial models to our domain.
Fine-tune and deploy these models into production applications across the business, partnering closely with product, engineering, and analytics.
Identify and mitigate risks like model bias and unreliable outputs to run foundation models reliably in a regulated financial product.
Design and optimize search and retrieval systems to improve discoverability, relevance, and user satisfaction.
Qualification
Comfort operating independentlyFluency in collaborating with ProductDrive and proactivity
Required
5+ years of experience building and productionizing ML/AI models that deliver measurable business impact.
Hands-on experience either pretraining foundation models from scratch (including large-scale, multi-GPU training) or fine-tuning LLMs for a specific domain, with a solid command of transformer architectures. Depth in one is enough; experience in both is a plus.
Solid grounding in traditional ML across use cases such as credit, fraud, ranking, recommendations, and personalization.
Comfort operating independently: you can scope ambiguous problems, figure things out with minimal direction, and know where to focus experimentation for maximum value.
A strong technical background with a degree in Computer Science, Data Science, Applied Mathematics, or a related field.
Fluency in collaborating with Product, Engineering, and Analytics teams to bring AI solutions from ideation to deployment.
Drive and proactivity: everyone here is a builder and executor.