How Hexploits engineered an AI-powered career intelligence platform for its clients

When a major charity in the education sector found itself facing a barrier of growing tech-debt, it turned to Hexploits for support. The consultancy designed a new system, while partnering to offer managed services that meant the charity can focus on its impact, rather than fighting fires in its technology operations.

Gradvisor is a UK-based charity on a mission to bridge the opportunity gap for students and recent graduates. The digital platform connects young people with over 600 professionals sharing authentic career insights across more than 80 career paths. For students without established networks, Gradvisor is vital, as a digital library of real experiences from relatable voices.

However, their technology had become an obstacle to their mission. As a charity operating on limited funding, the organisation had made early compromises to get to market quickly, but those compromises had compounded into a platform that was actively hindering operations – with the original engineering team having built a platform that, while functional, was buckling under its own weight.

Every week of delayed features was a week of missed impact – and amid this state of crisis, Gradvisor approached Hexploits. Founded in 2018, and based in London, Hexploits, is a consulting firm which helps clients to optimise their DevOps practices, migrate to the cloud, and ensure effective software releases – partnering at every point of the process.

At the time of engagement, user engagement was suffering at Gradvisor, according to an account of the project published by Hexploits. Page loads “averaged 10 seconds”, while the digital-native students with “zero patience for lag” were giving up on the platform before they could access the career insights they needed. With the team stuck firefighting on this basis, the external expertise of Hexploits presented a lifeline.

According to Hexploits, “We proposed what many consultancies avoid: a complete platform rebuild. Not because it was easy, but because it was the only path that served Gradvisor’s actual needs.”

Business objective

Hexploits determined the client needed a platform that would restore user engagement by delivering the instant responsiveness students expect; unblock feature development so they could iterate on their product and serve users better; and provide cost predictability so they could plan their charitable operations with confidence. This wasn’t a small undertaking, with details including 170 API endpoints being redesigned, rebuilt, and documented; and over 50 database tables to migrated from DynamoDB to a properly normalised relational schema, among other challenges. But “every endpoint was also an opportunity to fix the architectural mistakes of the past”.

In phase one of the process, Hexploits first designed the entire system from the ground up. A new relational database schema that actually modelled their domain. Clean API contracts. A deployment architecture built for performance and maintainability.

Then, the team rebuilt the platform on enterprise-grade technology—battle-tested frameworks chosen for long-term support and developer familiarity. This wasn’t about chasing trends; it was about building something any engineer could understand, maintain, and extend. A third phase meanwhile meant that, as Gradvisor’s AI and recommendation capabilities are what differentiate the organisation from a static content library, Hexploits architected dedicated services for these critical functions, designed to be cost-effective while delivering genuine intelligence.

Fitting into that, was a recommendation engine, which matches students with relevant career paths, job opportunities, sectors, and even other users with similar interests. An AI-powered career assistant was also included, with an intelligent chat agent with deep contextual awareness. Using retrieval-augmented generation (RAG), the assistant has comprehensive knowledge of the user’s profile, interests, and career journey. And a voice processing pipeline aimed to make the platform more accessible, with AI-powered transcription converting audio content to searchable text.

This all led to the fourth phase of the project.

“We introduced the practices that separate prototypes from production systems: comprehensive test suites: Unit tests and integration tests across the platform; infrastructure as code: Reproducible, version-controlled environments; proper release cycles; CI/CD pipelines: every change tested before it reaches users; and migration to Hexploits Cloud – as after stabilising the platform, we migrated Gradvisor to Hexploits Cloud – our managed infrastructure offering built for organisations that need enterprise reliability without enterprise complexity.”

Same infrastructure, better economics

Gradvisor now runs on the same underlying cloud infrastructure they had before, but with significant improvements. The organisation enjoys a cost reduction of 12% savings on equivalent AWS fees, and predictable billing, with fixed monthly costs, and no surprise bills. Meanwhile, it has access to faster CPUs and increased RAM at no extra cost, with managed operations – including deployments, rollouts, and maintenance – handled by Hexploits.

For a charity, every penny counts. But what matters even more is that Gradvisor is now freed up to focus on its core mission, rather than maintenance.

The consultancy concluded, “A charity lives and dies by its ability to demonstrate impact. Gradvisor needed more than a working platform – they needed the data to prove their value to funders, identify what’s working for students, and make informed decisions about where to invest limited resources. More importantly: students who would have bounced after 10 seconds of loading now stay, explore, and find the career insights they need. The technology no longer stands between Gradvisor and the people they’re trying to help.”