Machine Learning Engineer

Teleshop is an international trading and product sourcing company working across suppliers, manufacturers, products, customers, and global markets.

We are building intelligent systems to improve how trading companies manage product information, identify opportunities, evaluate suppliers, match products with customers, automate document processing, and support sourcing and commercial decisions.

We are looking for an Applied Machine Learning Engineer who can translate real operational problems into practical machine learning systems. The role will involve working with complex, incomplete, and highly varied trade data, including product catalogues, supplier documents, quotations, certificates, emails, customer interactions, and transaction records.

This is a broad, hands-on role covering machine learning, graph-based systems, natural language processing, document intelligence, recommendation systems, and production deployment.

• Design, build, train, evaluate, and deploy machine learning models for real-world trading and supply-chain applications.
• Develop systems for product, customer, supplier, and opportunity matching.
• Build recommendation and ranking models to support product selection, supplier sourcing, and customer targeting.
• Work with structured, semi-structured, and unstructured data from internal systems, documents, emails, catalogues, quotations, certificates, and product specifications.
• Develop natural language processing and document intelligence pipelines for entity extraction, relation extraction, classification, summarisation, and semantic search.
• Build and maintain knowledge graphs representing relationships between products, suppliers, customers, categories, transactions, and operational events.
• Apply graph-based machine learning techniques such as graph embeddings, link prediction, heterogeneous graph modelling, and graph neural networks.
• Develop entity-resolution and record-linkage systems for identifying duplicate or related companies, products, contacts, and documents.
• Work with large language models, embeddings, retrieval pipelines, structured outputs, and model-validation frameworks.
• Design models that can operate effectively with noisy, incomplete, weakly labelled, and limited datasets.
• Develop anomaly-detection, risk-scoring, forecasting, and decision-support models.
• Conduct feature engineering, model comparison, error analysis, explainability analysis, and uncertainty estimation.
• Build reliable data pipelines and APIs for integrating machine learning systems into existing business workflows.
• Collaborate with commercial, sourcing, operations, and technology teams to define measurable machine learning problems.
• Review and implement relevant methods from recent machine learning research.
• Document experiments, assumptions, evaluation results, and production decisions.
• Contribute to model monitoring, versioning, testing, and continuous improvement.

• Strong proficiency in Python.
• Strong experience with machine learning libraries such as PyTorch, TensorFlow, scikit-learn, pandas, and NumPy.
• Practical experience building, training, evaluating, and deploying machine learning models.
• Experience working with structured, semi-structured, and unstructured data.
• Knowledge of graph-based machine learning, including knowledge graphs, heterogeneous graphs, graph embeddings, link prediction, or graph neural networks.
• Experience with natural language processing and document intelligence, including entity extraction, relation extraction, document classification, and semantic search.
• Experience working with large language models, embeddings, structured outputs, retrieval pipelines, and model validation.
• Strong SQL skills and experience working with relational, document, vector, or graph databases.
• Ability to design robust data pipelines for sparse, noisy, incomplete, and weakly labelled datasets.
• Experience with model evaluation, feature engineering, error analysis, explainability, and uncertainty estimation.
• Familiarity with REST APIs, Docker, Git, and production machine learning workflows.
• Strong analytical and problem-solving skills.
• Ability to translate business and operational problems into measurable machine learning tasks.
• Ability to work independently in an environment where requirements may evolve as systems are tested and deployed.

• Experience with PyTorch Geometric, DGL, NetworkX, Neo4j, or related graph technologies.
• Experience with knowledge-graph construction, entity resolution, record linkage, and relationship inference.
• Experience extracting information from PDFs, supplier documents, certificates, audit reports, catalogues, purchase orders, quotations, emails, or product specifications.
• Experience with temporal graphs, time-series models, anomaly detection, risk modelling, or disruption propagation.
• Experience with candidate matching, recommendation systems, learning-to-rank, constrained ranking, or decision-support systems.
• Familiarity with optimisation tools such as OR-Tools, CVXPY, Pyomo, or Gurobi.
• Experience with cloud platforms such as AWS, Google Cloud, or Microsoft Azure.
• Experience with MLOps tools, experiment tracking, model versioning, monitoring, and reproducible pipelines.
• Knowledge of supply-chain management, procurement, manufacturing, logistics, international trade, supplier risk, or sourcing.
• Experience implementing or evaluating methods from recent machine learning research.

• Bachelor’s or Master’s degree in Computer Science, Machine Learning, Data Science, Artificial Intelligence, Statistics, Operations Research, or a related discipline.
• Relevant practical experience may be considered in place of a specific degree requirement.
• Prior experience delivering machine learning systems in a commercial, operational, or production environment is preferred.

Examples of potential projects include:

• Product and supplier recommendation systems.
• Customer-product matching and opportunity ranking.
• Product-category classification using text and image data.
• Knowledge graphs connecting suppliers, products, customers, documents, and transactions.
• Automated extraction of information from catalogues, certificates, quotations, and supplier documents.
• Supplier and counterparty risk scoring.
• Entity resolution across inconsistent company, product, and contact records.
• Semantic product search using text, images, and structured attributes.
• Forecasting and anomaly detection across sourcing and trading workflows.
• LLM-assisted systems for processing emails, documents, and operational records.

• Work on machine learning problems grounded in real international trade operations.
• Access a wide variety of commercial, product, supplier, and document data.
• Build systems that will be directly used by sourcing, sales, operations, and management teams.
• Take ownership of projects from experimentation through deployment.
• Work across multiple areas of applied machine learning rather than being limited to one narrow specialisation.
• Help shape the company’s machine learning infrastructure, research direction, and long-term product capabilities.
• For LinkedIn, I would keep the title as Applied Machine Learning Engineer and set the seniority level based on the experience requirement rather than adding “Senior” to the title unless you require at least four to six years of relevant experience.