Forward-Deployed Architect, Credit Analytics & ML Platforms

Neurons Lab LTD.; United States; À distance; USD 140 000 - 210 000; Plein temps

Résumé du poste

Neurons Lab LTD. seeks a hands-on Forward Deployed Engineering Architect to design and deliver a production-grade credit analytics platform for a regulated financial-services business operating in the UK and Ireland.

You will own the engineering architecture, contribute production code, and collaborate with the client’s business and technology teams to meet evolving portfolio needs. You will translate business problems into architecture, build reproducible ML workflows, ensure governance, and

Qualifications

• Production data or ML systems ownership beyond prototypes and diagrams.
• Strong hands-on Python and SQL; able to implement parts of solution.
• Practical MLOps: reproducible pipelines, versioning, testing, deployment and monitoring.
• Experience with cloud data platforms and data warehouses; Snowflake is highly relevant.
• Familiar with model evaluation, calibration, data leakage and reliability topics.
• Ability to navigate incomplete docs and translate into maintainable code and clear docs.
• Strong client-facing English; explain trade-offs to non-technical stakeholders.

Responsabilités

• Translate business problems into architecture, plan and acceptance criteria.
• Design production-ready ML platform and end-to-end workflows.
• Lead technical discussions, review code and unblock engineers.
• Ensure data governance, privacy, access control and model governance.
• Collaborate with modelling lead on PD, LGD, ECL and related metrics.

Connaissances

Python

SQL

MLOps

Data modeling

Leadership

Cloud platforms

Machine learning

English communication

Outils

Snowflake

Streamlit

Azure ML

Description du poste

Описание:

We are looking for a hands-on Forward Deployed Engineering (FDE) Architect to help develop a credit analytics platform for a regulated financial-services business operating in the UK and Ireland. The platform brings together lending and servicing data to support portfolio analysis, credit-risk modelling and better business decisions. Current work covers probability of default, redemption and prepayment, expected credit loss, and interactive analytics. The next challenge is to turn this work into a reliable, maintainable platform that can support different portfolios and client needs. You will work directly with the client’s business and technology teams, alongside a credit-domain and modelling lead and other engineers. You will own the engineering architecture, contribute production code, and help the team decide what to build and how to deliver it. We need someone who can question assumptions, propose simpler alternatives and explain technical trade-offs clearly.

Задачи:

Translate business problems into an architecture, practical delivery plan and clear acceptance criteria. Identify gaps in requirements, data and dependencies before they become delivery problems. Design the path from working models and dashboards to a production-ready ML platform. Make deliberate choices about shared components, portfolio-specific configuration, model calibration and future client onboarding. Build and improve reproducible workflows for data preparation, feature engineering, training, validation, scoring and deployment. Establish clear versioning for code, data, models and calibration logic. Work with the modelling lead on how probability of default, prepayment, lifetime risk and loss estimates are calculated, tested and exposed to applications. Keep business definitions consistent across models, tables and dashboards. Strengthen the data foundation: trace results to source, investigate data-quality issues, document assumptions and build automated checks around important transformations. Improve the reliability of the Snowflake-backed analytics application, including deployment, testing, permissions and integration with model outputs. Put appropriate MLOps practices in place, including repeatable releases, monitoring, failure handling, rollback and operational documentation. Design for privacy, access control, auditability and model governance. Work with domain experts to make validation, stability and fairness checks practical and reviewable. Lead technical discussions, review code, unblock engineers and keep the client informed about progress, risks and architectural decisions. Assess where natural-language or LLM-assisted analytics would add value, without losing focus on trustworthy data, robust models and useful products.

Требования:

A track record of designing and delivering production data or ML systems, with ownership extending beyond prototypes and architecture diagrams. Strong hands-on Python and SQL. You should be comfortable investigating a dataset, reviewing modelling code and implementing critical parts of the solution yourself. Practical MLOps experience: reproducible pipelines, model and data versioning, testing, deployment, monitoring and the operational consequences of model updates. Experience with cloud data platforms and modern warehouses. Snowflake experience is particularly relevant; you should be able to reason about data access, performance, cost and deployment constraints. An understanding of model evaluation, calibration, data leakage, changing populations and the difference between model quality and system reliability. The ability to work through incomplete documentation and inconsistent business definitions, and turn findings into maintainable code and clear documentation. Strong client-facing communication in English. You can explain trade-offs to non-technical stakeholders, constructively challenge a requirement and take responsibility for the resulting decision. Experience leading a small technical team while remaining actively involved in implementation. Credit-risk, lending or servicing analytics, including PD, LGD, ECL, prepayment or portfolio stress testing. Regulated financial-services delivery and model governance. Streamlit, Azure ML, semantic data layers or analytics applications built on Snowflake. Configurable ML products serving multiple portfolios or clients. Evaluated natural-language analytics, text-to-SQL or LLM-assisted workflows. Current technical environment Snowflake, SQL, Python, Streamlit and an Azure ML environment. The team also uses AI-assisted development tools. You will help evolve the architecture based on the product’s needs rather than introduce a predetermined stack. What we would like to discuss A production ML or analytics system you personally helped build, including the architecture decisions and code you owned. How you would adapt a model and its calibration for a new portfolio while preserving reproducibility, access boundaries and validation. A time you challenged a business or technical assumption, proposed an alternative and delivered a better outcome.