ML Engineer - Catalysk (Climate Fintech)

Catalysk isn’t just another startup — it’s a paradigm shift. We’re building the world’s first Sustainability Score for Individuals — like a credit score, but for your climate footprint. By converting electricity, water, commute, and spending data into an actionable rating, we’re creating the foundation for green finance, insurance, and incentives. At population scale.

You’ll be part of an early-stage team where your work is visible, impactful, and globally relevant.

About the Role
We’re looking for a curious, analytical, and experienced ML Engineer to join our team and turn raw data — especially financial transactions and household consumption data — into actionable insights. As we’re building a first-of-its-kind product, we need people who like a challenge and can think creatively.

You will work closely with our existing teams across merchant categorisation, machine learning and analytics, contributing directly where your data, NLP or ML expertise is useful.

This is a hands-on role where you’ll own the project end-to-end — from data wrangling to modeling to communicating results.

Key Responsibilities

Merchant categorisation & NLP (primary focus)

• Lead the technical approach to improving transaction narration understanding, merchant/entity extraction and classification, working alongside and guiding the existing categorisation team
• Build and evaluate transformer-based and other models, and decide where rules, models, LLMs or human review fit best
• Define and maintain training and evaluation datasets, with accuracy tracked by segment
• Analyse classification errors and drive fixes in the data, features or models
• Improve merchant mappings, aliases and transaction intelligence
• Review the team's work, set evaluation standards, and help develop junior colleagues

Production and feedback loops

• Deploy and monitor models, tracking drift and confidence as data volumes and client numbers grow
• Turn human-review corrections into retraining data and measure the improvement

Data, analytics and new signals

• Provide reliable datasets and support analytics where needed, with scope expanding as the team grows
• Explore new transaction signals and datasets that could improve Catalysk's models

Who We’re Looking For

• 3-5 years building and shipping ML systems into production, ideally classification, entity matching, or record linkage on messy real-world data
• Strong Python and SQL, and comfort with the full loop: data prep, training, evaluation, deployment, monitoring
• Experience designing evaluation sets and measuring performance by segment, not just a single headline accuracy number
• Practical judgement about when to use rules, when to use a model, and when to send something to a human
• Ability to work with imperfect labels and turn reviewer corrections into better training data
• Experience working with BERT, LLM and Generative AI technologies — fine-tuning, prompt engineering, or integrating models via APIs (e.g., OpenAI API, Hugging Face Transformers).
• Strong communication skills — ability to explain technical concepts to non-technical stakeholders.

Bonus:

• Experience working with financial data and/or on sustainability.
• Familiarity with retrieval-augmented generation (RAG), embeddings, and vector databases for building intelligent applications.
• Experience with data visualization tools (e.g., Tableau, Power BI, or Plotly), cloud platforms (AWS / GCP), or productionizing models.
• Interest in climate or sustainability, though domain knowledge isn't required; we'll teach the methodology

Why You’ll Love This Role

• Real Impact – Work on meaningful problems — they’ll shape how people and institutions act on climate. At population-scale
• High Visibility – Your work will directly feed into products used by banks, regulators, and corporations.
• Ownership – You’ll have freedom to experiment and design impactful models
• Learning Curve – Exposure to cutting-edge climate science, financial data, and sustainability methodologies.