Master’s in Technical NLP Engineering and Engineering Assistants
Online master's programme of 1,900 hours in technical NLP engineering and engineering assistants: 12 modules covering computational linguistics, Transformer models, RAG over technical documentation, information extraction, fine-tuning, agents, LLMOps and AI governance under the EU AI Act, ISO/IEC 42001 and NIST AI RMF, with an assigned technical tutor.
Master’s in Technical NLP Engineering and Engineering Assistants — studying from Malta
Malta has positioned itself as a centre for digital innovation, with government support for artificial intelligence, data science and blockchain-related activity. The Malta Digital Innovation Authority oversees policy in these areas, supporting a community of software companies and international firms. Demand for professionals skilled in analytics, simulation and machine learning is growing across financial services, gaming, aviation and engineering. Employers range from established multinationals to start-ups, offering opportunities for both technical specialists and those who translate data into business decisions.
Fees for Malta: €9,930 (Euro), with instalments and employer sponsorship available.
Relevant bodies in Malta
Typical roles
Questions from Malta
Are there professional bodies in Malta that recognise data and AI qualifications?
Malta has no single regulated title for data scientists or AI specialists. The Chamber of Engineers regulates 'engineer' only for those engaged in regulated engineering work. Industry associations and certification providers may offer their own credentials. SEIUM qualifications are private and not accredited by a national body, so employers decide how much they value them.
Which Maltese employers hire for data, AI and analytics roles?
Malta's iGaming, financial services, fintech, insurance and public sector organisations are major users of data and analytics, and a number of technology firms are based on the island. Employers generally assess practical skills through portfolios, technical interviews and past projects, so pair your qualification with demonstrable work.
Can I complete this while working in Malta, and what language and time commitments apply?
Yes. The programmes are 100% online and taught in English, with an assigned technical tutor and a final project. No student visa is required to study from Malta. Because the study is asynchronous, time zone is rarely an issue, but you should budget hours realistically, from 300 for specialist courses up to 1,900 for master's. Ask about instalment fees and employer sponsorship.
Take the controls.
Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.
What you will be able to do
Design language-processing pipelines for technical documentation.
Build RAG assistants with citations traceable to standards and manuals.
Extract requirements, tables and bills of materials from unstructured documents.
Adapt language models to the domain through efficient fine-tuning.
Orchestrate agents that use calculation, CAD and PLM tools.
Evaluate the reliability, security and biases of assistants in production.
Deploy engineering assistants in compliance with the European AI Act.
Who it's for
- NLP engineers moving into technical documentation and engineering assistants.
- LLMOps engineers responsible for running language models in production.
- Developers of technical assistants for engineering teams.
- Technical knowledge management specialists.
- Generative AI solution architects.
- AI governance leads in industrial organisations.
- Engineering or science graduates looking to enter engineering and technical consultancy.
What you'll work with
- Regulation (EU) 2024/1689 (European Artificial Intelligence Act)
- ISO/IEC 42001:2023 (AI management systems)
- NIST AI RMF 1.0 (AI risk management framework)
- ISO/IEC 23894:2023 (AI risk management guidance)
- Regulation (EU) 2016/679 – GDPR (personal data protection)
- ISO/IEC 22989:2022 (AI concepts and terminology)
- ISO/IEC 27001:2022 (information security management)
- Hugging Face Transformers
- spaCy
- LangChain
- vLLM
- Qdrant
12 modules · 1900 h
1.1Tokenisation of technical documentation and units1.2Morphology, lemmatisation and specialised terminology1.3Dependency parsing of written requirements1.4Named entity recognition in standards and data sheets1.5Vector representations: word2vec, fastText and GloVe1.6Technical corpora: manuals, specifications and tickets1.7Corpus annotation and inter-annotator agreement1.8Rule-based extraction and regular expressions1.9Evaluation metrics: precision, recall and F11.10Lab: spaCy pipeline on maintenance manuals
2.1Attention mechanism and multi-head self-attention2.2BERT-style encoders and GPT-style decoders2.3Masked and causal pre-training objectives2.4BPE, WordPiece and SentencePiece tokenisers2.5Context window and RoPE positional encoding2.6Scaling laws and choice of model size2.7Open models: Llama, Mistral and Qwen2.8GPTQ and AWQ quantisation and the GGUF format2.9Efficient inference with vLLM and KV-cache2.10Practical: serving an open model on a local GPU
3.1Zero-shot, few-shot and step-by-step reasoning prompting3.2System messages and style control3.3JSON outputs validated with Pydantic schemas3.4Function calling and tool use3.5Handling units and quantities in responses3.6Reducing hallucinations with explicit constraints3.7Prompt versioning and regression testing3.8Prompt injection and basic defences3.9Cost per token and latency optimisation3.10Workshop: requirements extractor to validated JSON
4.1Ingestion of PDFs, drawings and data sheets4.2Semantic and document-structure-based chunking4.3Embedding models and domain-based selection4.4Vector databases: FAISS, Qdrant and pgvector4.5Hybrid BM25 search with dense retrieval4.6Passage re-ranking with cross-encoders4.7Citations traceable to clause and page4.8Incremental updating of the document index4.9RAG evaluation with Ragas: faithfulness and relevance4.10Lab: RAG assistant over internal standards
5.1Requirements classification with EARS syntax5.2PDF tables: Camelot, pdfplumber and OCR5.3Reading title blocks and bills of materials5.4Normalisation of physical units with Pint5.5Detecting ambiguity in technical specifications5.6Relationships between components and knowledge graphs5.7Technical ontologies and controlled vocabularies5.8Linking to PLM part numbers5.9Human validation and correction loops5.10Case study: structured BOM from scanned specifications
6.1When to fine-tune the model and when to retrieve6.2Instruction datasets from technical reports6.3LoRA and QLoRA adapters with PEFT6.4Supervised fine-tuning and chat templates6.5Preference alignment through DPO6.6Synthetic data and contamination control6.7Catastrophic forgetting and regression evaluation6.8Distillation into small deployable models6.9Experiment tracking with MLflow6.10Practical: QLoRA on incident reports
7.1ReAct and planner-executor patterns7.2Integration with calculators and numerical solvers7.3Generated SQL queries over technical databases7.4Connection to CAD and PLM APIs7.5MCP protocol for exposing tools7.6Conversation memory and task state7.7Autonomy limits and human approval7.8Tracing and observability with Langfuse7.9Error handling and tool retries7.10Workshop: agent that sizes and documents a component
8.1Generating Python scripts for CAE pre-processing8.2Macros for FreeCAD and CAD APIs8.3PLC programming assistance according to IEC 61131-38.4Generated unit tests and code review8.5Static analysis and detection of insecure code8.6Technical documentation generated from code8.7Code licences and intellectual property8.8Integration into IDEs and review workflows8.9Productivity and code-quality metrics8.10Practical: automating meshing with an LLM
9.1Proprietary domain evaluation sets9.2LLM-as-a-judge and its biases9.3Confidence calibration and controlled abstention9.4Adversarial testing and red teaming9.5Data leakage and document privacy9.6OWASP Top 10 risks for LLMs9.7Risk management according to NIST AI RMF9.8Drift monitoring in production9.9Traceability for regulated sectors9.10Case study: audit of a maintenance assistant
10.1Cloud versus on-premise architecture10.2Containers, Kubernetes and GPU scaling10.3Model gateways and cost-based routing10.4Semantic caching and latency control10.5Secrets management and access control10.6Canary deployment and A/B testing10.7Audit logging of interactions10.8Total cost of operation of the assistant10.9Energy consumption of inference10.10Lab: deployment with vLLM and Kubernetes
11.1The European AI Act and risk levels11.2AI management systems according to ISO/IEC 4200111.3Personal data protection in assistants11.4Copyright of standards and manuals11.5Interaction design and user trust11.6Change management in technical offices11.7AI literacy for engineers11.8Return on investment and business case11.9Model cards and system documentation11.10Case study: implementation plan in an engineering firm
12.1Definition of the use case and scope12.2Collection and licensing of the technical corpus12.3Architecture design of the assistant12.4Implementation of the functional prototype12.5Evaluation plan and success metrics12.6Testing with real engineering users12.7Risk analysis and regulatory compliance12.8Code documentation and reproducibility12.9Master's Thesis report12.10Master's Thesis: defence before a panel
How you'll learn
100% online training on the SEIUM campus, with downloadable material, recorded classes and an assigned technical tutor. There are no fixed timetables: you progress at your own pace within each intake, so the programme is compatible with work.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is Hugging Face Transformers, spaCy, LangChain, vLLM and Qdrant, in free, academic or evaluation versions. Tests are studied on real procedures: evaluation of assistants on real technical corpora, trials with engineers in technical offices and auditing of responses against the cited standards. Assessment is continuous by module, plus the final Master's Thesis.
Standards covered: Regulation (EU) 2024/1689 on Artificial Intelligence, ISO/IEC 42001:2023, NIST AI RMF 1.0, ISO/IEC 23894:2023, Regulation (EU) 2016/679 (GDPR), ISO/IEC 22989:2022, ISO/IEC 27001:2022.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Where this takes you
Requirements
- A university degree in engineering, science or architecture, or equivalent demonstrable professional experience.
- An internet connection and a computer capable of running the programme software.
- English sufficient to read technical standards.
Ready to start?
Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.
Frequently asked questions
The programme totals 1,900 hours over 12–14 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.
You work with Hugging Face Transformers, spaCy, LangChain, vLLM and Qdrant, using free, academic or evaluation versions. Your tutor indicates the free alternative wherever one exists.
A SEIUM Institute diploma, awarded on passing the continuous assessment and the Master's Thesis. It is SEIUM's own qualification, not an official state degree; its recognition depends on the programme completed and the regulations of each country.
Yes. Modules passed in a course or postgraduate diploma in the same area are recognised in the higher programme. Ask about the recognition policy when you enrol.
A university degree in engineering, science or architecture (or equivalent professional experience), a computer able to run the software and English sufficient to read technical standards.
Start dates are set by SEIUM's faculty for each intake. Our admissions team will confirm the next available date when you apply.
Related programmes
Master’s in Time Series, Signals and Anomaly Detection
Online master's programme of 1,900 hours in time series, signals and anomaly detection: sensors, spectra, forecasting and root cause analysis. 12 modules covering digital signal processing,…
Master’s in Verification, Validation and UQ of Models
Online master's programme of 1,900 hours in verification, validation and uncertainty quantification of models: uncertainty, V&V, safety cases and confidence. 12 modules covering model credibility, code…
Master’s in Federated Learning & Privacy-Preserving ML
Online master's programme of 1,900 hours in federated learning and privacy-preserving ML: FL, differential privacy, encryption and multi-party governance. 12 modules covering federated algorithms, differential privacy,…
More in Data, AI, Simulation & Analytics
All Data, AI, Simulation & Analytics programmes → · How to apply · Fees & financing · Scholarships