A five-student team from Loughborough University has taken top honors at the Microsoft Embrace x Midlands Hackathon 2026 with Lighthouse, an AI-driven career guidance platform built in just five hours on Microsoft Azure. The rapid-prototyping victory, announced on May 11 at the university’s campus, showcased how cloud and AI tools can be harnessed to tackle a pressing real-world problem: helping young people navigate the increasingly complex maze of career options. The one-day hackathon, jointly organized by Microsoft and regional tech partners, challenged teams to build functional applications that leverage Azure AI services to address education, sustainability, or community challenges.

Inside Lighthouse: Real-Time Career Intelligence

Lighthouse addresses a perennial pain point for students and early-career professionals—turning a muddle of interests, skills, and market noise into a clear, actionable career path. The platform analyzes a user’s academic background, extracurricular activities, and personal preferences, then cross-references that profile against live labor-market data. It returns tailored job role suggestions, skill-gap analyses, and personalized learning roadmaps, all generated on the fly.

The interface, a clean web app optimized for mobile, guides users through a conversational intake. A natural-language processing engine deciphers free-text responses about hobbies and ambitions, while a recommendation model—likely built on Azure Machine Learning—weights factors such as salary expectations, regional demand, and automation risk. During their live demo, the team showed how a second-year psychology student could be matched not only to clinical roles but also to emerging fields like UX research or AI ethics, complete with instant links to relevant courses and entry-level job listings.

Five Hours, One Mission

The hackathon’s format was deliberately intense. At 9:00 a.m., organizers unveiled three challenge themes: Future of Work, Sustainable Cities, and Digital Inclusion. Teams had until 2:00 p.m. to submit a working prototype, a pitch deck, and a short demo video. Industry mentors roamed the floor, but the builds had to be student-driven. The Loughborough quintet—three computer science majors, one design student, and one business undergraduate—split tasks along natural lines: two focused on the Azure backend, one on the React frontend, one on data wrangling from public APIs, and the designer polished the user journey.

By 11:30 a.m., the group had a rough pipeline feeding test data into Azure AI Search and Azure OpenAI Service. By 1:00 p.m., they were refining the prompt engineering that makes Lighthouse feel conversational rather than mechanical. “We wanted it to ask smart follow-ups, not just spit out a list,” one team member explained during the presentation. The final minutes were spent stress-testing the recommendation logic against edge cases—what if a user has zero formal qualifications but years of volunteering?

The Tech Stack That Powered the Win

Lighthouse’s architecture shows how much can be accomplished with the Azure ecosystem under a tight deadline. The back end relies on Azure Functions for serverless compute, ensuring the app can scale from zero users to thousands without manual intervention. User profiles and job-market snapshots sit in Cosmos DB, whose multi-model flexibility allowed the team to store both structured skill matrices and unstructured chat histories.

At the core sits Azure AI Foundry, the unified platform that let the team chain together pre-built models. A custom Language Understanding (LUIS) intent recognizer picks apart a user’s “I like working with people but I also want a high salary” into structured preferences. Those preferences flow into a semantic ranking algorithm that queries an indexed corpus of career data—sourced from government labor statistics and real-time job boards—via Azure AI Search. The ranking model was fine-tuned in minutes using a small set of example profiles, a task that once would have taken days.

A particularly clever touch was the integration of Azure Maps. For each recommended career, Lighthouse displays a heatmap of job availability within a commute radius, pulling location data from open datasets. This visual cue helps users ground abstract job titles in real geographies, making the output feel tangible rather than theoretical.

The team also connected the platform to Microsoft Teams, generating a shareable summary card so that users can discuss their options with parents or career counselors. That level of polish—in five hours—sealed the win.

Why This Matters Now

Career indecision is not a minor inconvenience. A 2025 survey by the UK’s Career Development Institute found that 68% of university applicants felt overwhelmed by the number of choices, and 41% changed their intended course after accepting an offer. Mismatch between qualifications and labor-market needs costs the UK economy an estimated £6.2 billion annually in retraining and lost productivity. Tools like Lighthouse promise to close that information gap before it widens, giving young people data-driven confidence at a fraction of the cost of private career coaching.

Microsoft’s own education strategy has been doubling down on AI-assisted career readiness. The company’s Career Coach, a Microsoft Teams app rolled out in 2025, already offers skill assessment and course recommendations, but Lighthouse extends that vision by adding external labor-market signals and generative AI conversation. The hackathon entry effectively prototypes a next-generation version of those tools, built with the same cloud services that enterprise customers use.

Competition and Judging

Seven university teams participated in the hackathon, with entries ranging from a carbon-footprint tracker using Azure IoT to an accessibility plugin for virtual reality classrooms. The judging panel—comprising Microsoft cloud solution architects, local startup founders, and a careers adviser from Loughborough’s own award-winning student services—evaluated submissions on technical execution, innovation, social impact, and presentation clarity.

Lighthouse scored near-perfect marks on impact. One judge noted that the platform “turned what is usually a six-month soul-searching exercise into a 10-minute interactive session.” The runner-up, a water-leak-detection system for city utilities, impressed on technical grounds but was seen as less immediately usable by non-experts. Several judges remarked that Lighthouse could be productized with relatively little extra work, which likely tipped the balance.

What Comes Next

The Loughborough team plans to continue development. Microsoft has offered the group Azure credits and ongoing mentorship through the Microsoft for Startups Founders Hub, a program that already supports student-led ventures from concept to commercialization. Near-term goals include adding a resume-parsing feature that extracts skills from PDFs and a voice interface so that users can talk through their career dilemmas while, say, commuting.

Talks are underway with the university’s careers service to pilot Lighthouse during the autumn term. If that trial succeeds, the platform could become a permanent feature of the student portal, potentially integrated with the existing Microsoft 365 and LinkedIn Learning environments that Loughborough already licenses. The team is also exploring open-sourcing the career-data ingestion pipeline, which would let career offices anywhere plug in local labor statistics.

AI Hackathons as Talent Incubators

Microsoft’s Embrace series, which now includes regional events across the UK, Ireland, and India, is part of a broader industry trend: hackathons are no longer just about t-shirts and pizza. They serve as live auditions for employers. Four of the five Lighthouse team members have already been contacted by recruiters from Azure-focused consultancies. “Students who can build real-world AI applications in half a day are exactly who we want,” an engineering manager at a Microsoft partner firm said, speaking on condition of anonymity because the firm had not yet formalized offers.

For Microsoft, the return on investment is dual: the company seeds its ecosystem with fresh applications that demonstrate Azure’s capabilities, while simultaneously identifying future hires who are comfortable with low-code and AI-native workflows. The Embrace x Midlands event even included a mini “tech talk” session where Azure engineers demonstrated new features from the Build 2026 conference, turning the hackathon into a two-way learning street.

The Broader Context: AI in Education

The win comes as universities wrestle with AI’s role in teaching and assessment. While some institutions have banned generative AI over plagiarism fears, others are embedding it into the curriculum. Loughborough University has taken the latter path, recently launching a mandatory AI literacy module for all first-year students. Lighthouse demonstrates the potential upside: using AI not to replace human guidance but to scale it, so that every student can access personalized career advice—not just those who can afford private counselors.

The UK government’s 2025 AI Opportunities Action Plan explicitly calls for “AI-powered career navigation tools” to be made available in secondary schools and further education colleges. A project like Lighthouse, if it evolves beyond prototype, could align with those public-policy goals. The team says it is already speaking with a local further education college about adapting the platform for 16–18-year-olds.

A Five-Hour Project with Long-Term Promise

The Lighthouse project proves that ambitious, socially useful AI tools don’t always require months of development. With modern cloud platforms, a small, multidisciplinary team can go from idea to working demo in a working day. The real test will be whether the team can sustain momentum after the hackathon glow fades—turning a clever prototype into a maintained, secure, and inclusive service. If they succeed, Lighthouse might become more than a hackathon winner; it could change how a generation thinks about work.