A Decade of Data Science: How Our Programme Has Evolved Since 2014

If you look back at the data science landscape in 2014, it feels like a

different world. We were building churn models on personal laptops and sharing datasets over email. Today, we are helping our partners deploy bespoke solutions wrapped around Large Language Models (LLMs) and advanced predictive tools in secure, reproducible cloud environments.

To ensure our enterprise and portfolio partners receive rapid, cutting-edge AI solutions without disrupting their core engineering teams, our programme has constantly adapted. Across 35 cohorts, we have completely transformed our tech stack, our agile delivery models, and our project oversight. Here is a look at how S2DS has evolved to meet the demands of modern business.

The Tech Stack: From Laptops to High-Compute GPUs

The tools our teams use to build your commercial solutions have evolved rapidly to integrate seamlessly with modern tech stacks, follow best practices and ensuring secure high-standard delivery.

2014

Built solutions around classical ML (like classification and regression) in both Python and R; local development done on personal laptops across many platforms; data shared via email.

2020-2023

Migrated to secure cloud infrastructure (AWS Workspaces); standardised on Python; projects often required tech like NLP and early transformers models or image processing and neural networks.

2026

Provide GitHub Codespaces for secure and standardised cloud compute; projects accelerating with GenAI assistance; rigorous benchmarking of GenAI vs. classical ML to deliver business ROI.

Delivery Methodology: From Lecture Halls to Agile Execution

We realised early on that to deliver tangible ROI in just five weeks, our data scientists needed to operate precisely like your internal commercial tech teams, hitting the ground running from day one.

2014

Up-front lecture-heavy training (ML/data vis) before project work; organic, ad hoc planning.

2020-2023

Shifted to a Project First model with self-directed pre-learning, interactive workshops, and formal Agile methodologies (scrums, sprints) to ensure rapid prototyping.

2026

Fully integrated Agile lifecycle, embedded commercial AI literacy. Refined via feedback from over 35 cohorts to guarantee predictable, high-quality project delivery.

Commercial Oversight: From Pub Quizzes to Elite Talent Pipelines

Delivering advanced AI solutions rapidly requires expert technical oversight. We updated our mentorship model so your core team doesn't have to manage development sprints, while simultaneously building a talent pipeline for your future hiring needs.

2014

Centralised, informal mentorship; community built through in-person social bonding; nurtured a diverse global network by running an online programme each year.

2020-2023

Shifted online and built a mentor ecosystem (Technical, Team, Partner) to better align with client stakeholders; promoted networking with online activities (pub quiz!) and async communication platforms (like Slack, Teams).

2026

Focus on commercial AI (ensuring solutions are reproducible, ethical, and cost-effective); cultivate an elite pool of Fellows with 1-to-1 coaching and job board access; recruitment options from our talent pipeline.

A lot has changed over the years, new tech stacks, new industry challenges, and new delivery formats. But the one constant has been the calibre of the people. Looking back at how S2DS has grown, our greatest achievement isn't just the projects we've delivered, but the thriving, global alumni network we’ve built.