Company

Pointellis, developed in partnership with EY, tackles one of healthcare's most critical challenges: helping oncologists make personalized treatment decisions. The platform synthesizes patient genomic data, clinical histories, and real-world outcomes to recommend optimal treatment pathways.

The Challenge

Healthcare generates massive amounts of data—80% of it unstructured and 97% going unanalyzed or unused (Source: PLOS Digit Health/Cornell University). Pointellis needed to transform this chaos into clarity, building machine learning pipelines that could process genomic sequences, clinical records, and treatment outcomes into actionable insights. The technical complexity was significant: strict HIPAA compliance, limited labeled training data, and the need for complete transparency in AI recommendations. For oncologists to trust the platform with life-and-death decisions, it needed to be both clinically accurate and fully explainable.

The Solution

As Product Lead, Maheep Bhalla, a member of A.Team's network, brought rare expertise bridging ML capabilities with clinical requirements. His understanding of both healthcare workflows and AI limitations enabled him to build a platform oncologists would trust.

Bhalla orchestrated collaboration between oncology experts, data scientists, and engineers to translate complex requirements into actionable features. The team implemented proxy-labeling techniques and retrospective EMR validation to overcome limited training data. He prioritized transparency in the physician dashboards, ensuring doctors could understand exactly why the system recommended specific treatments. Through pilots at two U.S. cancer centers, the team refined both models and UX based on real-world clinical feedback, creating a system that balanced sophistication with usability.