DR-03 — AI Tools for Coding & Development
Purpose: Profile AI coding assistants and computational research tools for University of Sheffield researchers. Only MATLAB Copilot is currently catalogued; GitHub Copilot, Google Colab, and others mentioned in the intelligence report are absent.
Use with: Claude Research mode (or equivalent deep research tool)
Prompt
You are helping a UK research university (University of Sheffield) evaluate AI-assisted coding and computational research tools. Profile each tool below for institutional procurement and researcher recommendation.
For each tool, research and return:
- Primary function — one sentence
- Key AI features — code completion, generation, explanation, debugging, refactoring, etc.
- Languages and environments supported — programming languages, IDEs, notebooks
- Licensing and cost — including any free tiers for students/educators/academics
- Data training policy — does code or prompts get used to train models? Any enterprise/institutional opt-out?
- GDPR/data residency — EU/UK hosting? DPA available?
- SSO support — does it support SAML/institutional SSO?
- HPC/cluster compatibility — can it be used on university HPC systems or remote servers?
- Academic/HE adoption — named university deployments or case studies
- Catalogue status recommendation —
adopt | trial | assess | hold
Tools to profile
Free / zero-cost for academia:
- GitHub Copilot (Microsoft/GitHub) — AI code completion; free for verified students and educators
- Google Colab (Google) — cloud-based Jupyter notebooks with free GPU; zero setup
- Google Gemini Code Assist (formerly Duet AI) — AI coding assistant in Google Cloud / Colab
- Microsoft Copilot in VS Code — AI coding features in VS Code (free tier available)
Paid / institutional:
- Cursor — AI-native code editor (fork of VS Code)
- Tabnine — AI code completion with on-premises/private deployment option
- Amazon CodeWhisperer — AWS-integrated AI coding assistant (free individual tier)
- JetBrains AI Assistant — AI features across JetBrains IDEs (PyCharm, IntelliJ, etc.)
Open-source / self-hostable:
- Continue (continue.dev) — open-source AI coding assistant for VS Code / JetBrains
- Ollama + coding models — local LLM deployment for air-gapped or sensitive research code
Notebook environments:
- JupyterHub — multi-user Jupyter server (self-hosted); AI plugin ecosystem
- Deepnote — collaborative data science notebooks with AI features
- Databricks — unified analytics platform with AI/ML notebook environment
Additional research questions
-
HPC integration: Which of these tools can be used effectively when SSH-connected to a remote HPC cluster (e.g. Sheffield’s Stanage/Bessemer systems)? Are there VS Code Remote/SSH + Copilot combinations that work on HPC?
-
Sensitive code: What are the data governance implications of using cloud-based AI coding assistants when writing code that handles sensitive research data (health records, personal data)? Which tools offer on-premises or private-cloud deployment for this use case?
-
Research software engineering: Are there AI tools specifically designed for research software engineering workflows (FAIR software, reproducibility, software citation, dependency management)?
Output format
Return a markdown table:
| Tool | Primary function | AI features | Languages | Cost (academic) | Data training policy | GDPR | SSO | HPC-compatible | Recommended status |
|---|
Follow with a priority recommendation: which 2–3 tools should the University of Sheffield recommend immediately as zero-cost adopt options, and which should go to trial?
Include a note on any tools where on-premises deployment makes them suitable for sensitive/classified research computing.