Head of ML & MLOps Engineering - Fintech Engineering

InPost

About The Position

Build and lead the ML and MLOps function from the ground up within a scale-up inside InPost, as part of the core Data & AI leadership team. The role is accountable for creating a state-of-the-art ML platform and the discipline around it, with models tied to business cases and measurable monetary outcomes. It includes owning model outcomes from initial idea through profit-and-loss impact, and managing the function’s compute budget, headcount, and return on investment. The role leads development of credit and decisioning models with rigorous validation, champion/challenger testing, and explainability. It oversees a production ML platform where serving, monitoring, reproducibility, and retraining are engineered into the workflow. Responsible AI practices include model-risk, bias, and explainability checks, plus an independent sign-off gate before production. Agent-based systems are treated as production components with orchestration, guardrails, evaluations, and observability, built on a point-in-time-correct feature store. The position establishes model-development standards and validation methodology suitable for regulatory and model-risk scrutiny, leads agentic model iteration and monitoring, and owns the reliability, cost, and latency of production agent loops and graphs. It also defines the ownership boundary between feature production and model consumption with Data Engineering leadership and the Director. The role is hands-on when needed and collaborates with Data Engineering and Credit Risk.

Requirements

Experience across the full ML lifecycle: development, validation, deployment, monitoring, and retraining.
Experience architecting agentic systems for ML, including orchestration, guardrails, evaluations in CI, human-in-the-loop checkpoints, and agent observability.
Experience building and leading ML teams in regulated settings.
Experience in credit-scoring or underwriting modelling, or comparable high-stakes machine learning.
Skill in model-risk management and responsible-AI governance.
Experience building and leading a team from zero.
Fluent English (B2+).
Ability and willingness to be hands-on when needed, understanding models, pipelines, and serving layers.

Nice To Haves

Knowledge of CCD2 and consumer-credit regulation.
Knowledge of DORA/ICT risk.
Knowledge of IFRS 9 implications for model outputs.
Experience with fraud-detection machine learning.
Experience with Databricks/Spark.

Responsibilities

Build the ML and MLOps function and hire the team from the first hire onward.
Build and lead the ML platform for decisioning services, including serving, monitoring, reproducibility, and retraining.
Establish model-development standards and validation methodology that can withstand model-risk and regulatory scrutiny.
Own agentic model iteration, monitoring, and observability, bound to the feature store.
Own the reliability, cost, and latency of agent loops and graphs in production.
Ensure models have business cases and measurable monetary outcomes; manage the function’s compute budget, headcount, and return on investment.
Oversee credit and decisioning models with rigorous validation, champion/challenger testing, and explainability.
Ensure responsible-AI checks for model risk, bias, and explainability, including independent sign-off before production.
Build on a point-in-time-correct feature store and establish a clear ownership line between feature production and model consumption with Data Engineering leadership and the Director.
Architect agentic ML systems, including orchestration, guardrails, evaluations, human-in-the-loop checkpoints, and observability.