Pathful Reinvents Career Connections with AI-Powered Matching on AWS

How do students discover and pursue fulfilling careers? For Pathful, one of the answers is: Talk to a professional. The college and career readiness platform connects millions of students with real-world career experiences by facilitating live sessions with professionals working in the industries and roles students want to learn more about. These interactions help students see what’s possible and build confidence in their own career paths. With more students seeking deeper, more personal connections in a digital-first world, Pathful has seen the demand for this service rise to the point where the team grew concerned about its ability to scale effectively.

While scale was a challenge, so was quality. Educators were often too busy to provide detailed session requests, and professionals described their skills and roles inconsistently. This made it difficult to find the right match for the right student or class. And when the system failed to connect the two sides, the internal team at Pathful had to step in. “If our system couldn’t find a match, it went back to a human to do it, so there was a material impact,” said Paul Johansen, chief product and technology officer[SH1]  at Pathful. To fix these challenges, the company needed a better way to create efficient and effective matches, without losing the authenticity of the platform’s experience.

To reimagine its matching experience, Pathful partnered with Amazon Web Services (AWS) Premier Tier Services Partner Caylent to build a new backend designed for flexibility and efficiency at scale. At the core of the solution is a redesigned matching engine powered by natural language processing and AI services from AWS. The new approach uses generative AI to enrich and standardize data on both sides of the match: teachers’ session requests and professionals’ backgrounds. That way, the system can suggest more relevant and personalized connections without requiring extra input from users. Caylent worked closely with Pathful to design this AI-powered workflow, introducing tools to reduce friction while preserving the human-centered nature of each session.

Under the hood, the solution is built around Amazon API Gateway, which acts as the central entry point for backend services. API requests are authenticated using Amazon Cognito, where Caylent migrated all legacy professional accounts to a new identity store. Once authenticated, requests are processed by AWS Lambda functions that handle tasks like session retrieval, user updates, and match requests. The front end, powered by AWS Amplify, offers isolated environments for streamlined management and testing. “This wasn’t a typical implementation,” said Vinicius Defeo, senior architect at Caylent . “It combined infrastructure, software, and data engineering in a way that let us push the boundaries of what intelligent matching could look like.”

And to support that intelligent matching, Caylent integrated Amazon OpenSearch Service and Amazon Bedrock. Amazon OpenSearch powers full-text and semantic search capabilities to return high-quality results even with incomplete or inconsistent data. Amazon Bedrock contributes generative AI capabilities, enriching educator and professional profiles with standardized metadata. Together, these services form a modular, cloud-native foundation that supports future growth.

Pathful is preparing to roll out its new AI-powered matching solution and is confident that it will drive major improvements for students, educators, and its own internal teams. The new backend is projected to cut infrastructure costs by 70 percent and improve its ability to make a match to above 90 percent. The solution lays the technical foundation for more personalized experiences that benefit all three audiences: students, teachers, and the industry professionals who volunteer their time.

The work with Caylent and AWS is also helping Pathful think beyond matching. “We’re taking the core outputs of these live sessions—real people sharing real stories—and using AI to expand their reach,” said Johansen. “This is about helping students see themselves in these roles, and we’re building a foundation that will help us do that more meaningfully and at scale.” The team’s next initiative focuses on optimizing how live sessions are processed and shared—each one recorded and edited for asynchronous learning. Pathful plans to consolidate its video infrastructure, improve accessibility with multilingual transcripts, and generate rich metadata to support other search and recommendation features.

Additionally, this project has deepened Pathful’s relationship with Caylent and AWS. “For us, Caylent has become an essential partner that amplifies our team's capabilities, helping us accelerate innovation and deliver value to our students and educators faster than we could on our own,” said Johansen. “It’s a great partnership where Caylent works extremely well with AWS, and together they help us move our solutions forward.” With that combined support, Pathful is confidently building toward a future of smarter, more scalable student-professional connections.