Program Overview
This executive workshop is designed to help scientists and engineers transform how they conduct research, experimentation, and discovery using applied AI. It addresses the significant time lost in non-insight tasks such as literature review, data preparation, and reporting by introducing AI-driven workflows and tools. Through a structured Discover → Apply → Build → Deploy framework, participants gain hands-on experience with real datasets, enabling them to move from conceptual understanding to deploying production-ready AI solutions. The program emphasizes practical, no-code applications of AI tailored to industrial R&D environments.
Requirements
- No coding required — all exercises are no-code or natural-language first
- Basic familiarity with Excel and scientific datasets strongly preferred
- Participants encouraged to bring their own R&D dataset (optional)
- Laptop with Chrome browser; AI tool access pre- configured by faculty
Features
- Summarise literature, extract insights, and generate structured hypotheses using AI
- Perform data analysis, EDA, and feature engineering with no-code AI tools
- Build and interpret machine learning models without coding
- Apply AI-assisted Design of Experiments (DoE) to reduce experimental runs
- Develop AI workflows, automated pipelines, and deploy AI agents/web apps
- Automate code generation, debugging, and data analysis using AI
- Design and launch a 90-day AI pilot while ensuring AI safety, ethics, and governance
Target audiences
- R&D Scientists & Senior Researchers
- Process & Systems Engineers
- Lab Managers & Principal Scientists
- Innovation & Technology Leads
Curriculum
- 16 Sections
- 61 Lessons
- 3 Days
Expand all sectionsCollapse all sections
- Opening & AI Myth-Busting3
- How AI Works — Architecture & Key Terms3
- Prompt Engineering for Scientists4
- 3.1COSTAR Framework: Context, Objective, Style, Tone, Audience, Response format
- 3.2Role-prompting, chain-of-thought, few-shot and zero-shot with R&D examples
- 3.3Structured output prompts: extract data tables, mechanisms, and gaps from papers
- 3.4Hands-on: write 5 research-grade prompts on your own domain in real time
- AI for Literature Review & Knowledge Synthesis4
- 4.1ChatGPT / Claude: upload multi-paper sets — extract mechanisms, identify gaps
- 4.2Perplexity AI for real-time patent landscape and prior-art review
- 4.3NotebookLM: build a queryable R&D knowledge base from internal reports
- 4.4Reduce literature-to-hypothesis time from days to under one hour — live demo
- Data Analysis, EDA & Feature Engineering4
- 5.1Excel Copilot: natural-language pivot analysis, formula generation, data narration
- 5.2Python + Pandas via no-code interface: outlier detection and feature importance
- 5.3AI-assisted exploratory data analysis on real industrial process datasets
- 5.4Identifying hidden correlations in messy experimental and sensor datasets
- AI Safety, Ethics & Governance4
- 6.1Hallucination — how it happens, how to detect it, and how to design against it
- 6.2Data confidentiality: what must never be submitted to a public LLM
- 6.3Responsible AI in regulated R&D environments: GxP, ISO, SOX considerations
- 6.4reflection exercise — 90-second synthesis shared across full cohort
- AI Workflow Automation — Copilot Studio & Claude4
- 7.1Building multi-step AI workflows without writing a single line of code
- 7.2Copilot Studio: design an agent for lab documentation and reporting automation
- 7.3Claude Skills, Cowork, Code — capabilities, constraints and R&D use cases
- 7.4Connecting triggers, actions, and automated reporting pipelines end-to-end
- Design of Experiments (DoE) & Hypothesis Generation4
- 8.1AI-assisted DoE: fractional factorial designs — 70–80% fewer experimental runs
- 8.2Bayesian optimisation and active learning for iterative R&D experimentation
- 8.3JMP + AI Assistant: GUI-based response surface modelling with AI factor selection
- 8.4Group Activity: generate a DoE matrix for your own process problem using AI
- Build & Interpret Your First ML Model — No Coding4
- 9.1Upload a process dataset; instruct AI to train a regression model end-to-end
- 9.2Interpreting variable importance, RMSE and residual plots for non-coders
- 9.3DataRobot / AutoML: automated model training, cross- validation and deployment
- 9.4Communicating ML findings to non-technical stakeholders with confidence
- Process Optimisation — Industrial Deep Dives5
- 10.1Aluminium smelting: LSTM networks for current density and anode effect prediction
- 10.2Cement kiln: ML-driven heat balance, fuel optimisation and clinker phase forecasting
- 10.3Copper flotation: computer vision for froth analysis and reagent dosing control
- 10.4Chlor-alkali: survival analysis models for ion-exchange membrane life prediction
- 10.5Carbon black: particle size and morphology prediction from furnace operating data
- Chemistry, Materials & Molecular AI2
- Build & Deploy a Working AI Agent4
- 12.1Agent architecture: perceive → reason → act → report — the agentic loop explained
- 12.2Build a literature search and summarisation agent using Claude API or Gemini
- 12.3Deploy live to Streamlit Cloud — a shareable web app within 90 minutes
- 12.4Rate limiting, cost management, and secure API key practices for R&D teams
- R&D Code Automation with Claude Code4
- 13.1Generating Python / R analysis scripts using Claude Chat and Claude Code
- 13.2Debugging, refactoring, and documenting legacy analysis code with AI assistance
- 13.3Creating automated EDA pipelines: ingest → clean → visualise → narrate
- 13.4Hands-on: write, run, and modify a full data analysis script during the session
- Rapid Experimentation & Digital R&D Framework4
- 14.1The Intelligent R&D Loop: hypothesis → AI-design → run → learn → iterate
- 14.2Reducing experimental friction: predictive pre-screening via active learning
- 14.3Institutional knowledge capture: AI tools that retain expertise before it leaves
- 14.4Measuring AI impact: leading productivity indicators and transformation metrics
- Your 90-Day AI Pilot — Implementation Roadmap4
- 15.1Framework: Problem Statement → Data Source → Success Metric → 90-Day Plan
- 15.2Participant-led pilot scoping: one concrete, named use case per attendee
- 15.3AI governance checklist for R&D organisations: data, model, risk, and audit
- 15.4Briefing structure to present your AI pilot to leadership with confidence
- Post-Quiz, Capstone Showcase & Certification4
- 16.1Post-workshop knowledge benchmark — compare against Day 2 pre-quiz score
- 16.2Certificate of Completion awarded — Applied AI for R&D, Executive Programme
- 16.3Capstone showcase: 3-minute AI pilot pitch per participant to the full cohort
- 16.4Q&A, faculty feedback, peer networking, and 30-day follow-up commitments
Instructor
