Machine Learning Engineer (NLP)

About Us UMO is a stealth-mode FinTech venture aiming to evolve the way people experience money by building a unified, AI-powered, yet deeply human modern money platform across fiat, crypto, and investments - subject to regulatory approvals. The platform is being designed to break down traditional barriers to money across access, assets, and experience, enabling simpler, more adaptive ways for people to interact with financial services. We are currently developing our MVP and navigating licensing requirements, with a multidisciplinary team of 100+ people representing 20+ nationalities. With our headquarters in the UAE and offices in Portugal and Ukraine, we are united behind a shared ambition and a relentless focus on serving our customers. Key Responsibilities: Financial Sentiment Analysis: Build and deploy NLP models to analyze news, social media (Twitter/X, Discord), and Reddit to gauge market sentiment for stocks and crypto assets. Named Entity Recognition (NER): Develop systems to identify and extract entities (tickers, company names, wallet addresses, transaction IDs) from unstructured financial documents and chat logs. Automated Document Processing: Create pipelines to parse and extract data from financial statements, whitepapers, and regulatory filings (e.g., SEC filings) to assist in automated research. Fraud & Anomaly Detection: Implement NLP techniques to analyze transaction metadata and communication patterns to identify potential money laundering (AML) or fraudulent payment activity. Intelligent Customer Support: Build or fine-tune LLMs (Large Language Models) to power specialized chatbots capable of answering complex queries about portfolio performance, crypto protocols, or trading rules. Search & Discovery: Optimize internal search engines using semantic search and embeddings to help users find relevant financial instruments or transaction history. Model Lifecycle Management: Manage the full MLOps lifecycle, including data labeling for financial jargon, model training, deployment via APIs, and monitoring for "model drift" in volatile markets. Requirements: BA, Master’s or PhD in Computer Science, Data Science, or a related field with a focus on Natural Language Processing or Deep Learning. Advanced proficiency in Python and deep learning frameworks such as PyTorch or TensorFlow. Proven experience with Transformers (BERT, RoBERTa), Large Language Models (LLMs), and vector databases (e.g., Pinecone, Milvus, or Weaviate). Strong experience in building data pipelines using tools like Spark, Kafka, or Airflow, and proficiency in SQL. Familiarity with financial terminology and the ability to handle domain-specific data challenges (e.g., interpreting ticker symbols vs. common words). Experience deploying models in a cloud environment (AWS, GCP, or Azure) using Docker and Kubernetes, ensuring low-latency inference for real-time trading signals. Ability to design robust evaluation frameworks for NLP models, moving beyond standard metrics to business-impact metrics like "signal-to-noise ratio" in trading. A "builder" mindset with the ability to prototype rapidly and move from a research paper to a production-ready feature in weeks, not months. Fluent in English with excellent documentation and cross-team coordination skills The UMO Standard: Dynamic Work Environment: Vibrant offices and strong local teams across Lisbon, Dubai, Kyiv, and Lviv — with openness to remote for the right fit. Rest & Recovery: 24 days of annual leave, dedicated paid sick leave, and observance of Public Holidays to keep you at your best. Recharge Week: After your first year with us, enjoy two consecutive 4-day work weeks annually — a built-in reset to help you sustain peak performance for the long run. Full Setup: Top-of-the-line hardware and a home office stipend so you can do your best work anywhere. Growth Ecosystem: A dedicated learning budget and a clear, accelerated path toward executive-level leadership within the UMO ecosystem. Impact & Ownership: Your work directly shapes the product and company direction — not just executes it.