Cryptography · Compilers · Biomedical Informatics
I build systems that let organizations compute on data they cannot see.
Postdoctoral Associate at Yale, working on cryptographic protocols for biomedical and AI workflows. Author of Sequre and Shechi. Available for select engagements.
I use modern cryptographic techniques to enable secure computation on private data. My work spans the full stack of secure computation systems, from the design of new protocols to the implementation of high-performance compilers and frameworks for secure multiparty computation, homomorphic encryption, trusted execution environments, and federated learning.
Most privacy-enhancing computation projects fail due to inadequate feasibility study. I help teams scope their problem practically — assessing what's achievable with MPC, FHE, TEE, or federated approaches before engineering budget is committed.
Building production-ready secure computation systems on top of Sequre, Microsoft SEAL, Lattigo, and similar frameworks. Typical engagements cover the design, the prototype, and the handoff to the client's engineering team.
Training and inference on private data — regression, classical ML, neural networks of tractable scale, and federated approaches. Matching the right cryptographic technique to the model architecture and performance budget.
Technical advisory for engineering and research teams entering this domain. Workshops, code review, protocol selection, and long-term collaboration with research groups working at the intersection of cryptography and compilers.
Postdoctoral Associate at Yale University, working with Hyunghoon Cho. Leading an interdisciplinary effort across Yale, Stanford, and UCSD to deploy the first secure distributed workflow between the U.S. Department of Veterans Affairs and the National Institutes of Health for analyzing private genomic data.
Ph.D. in Computer Science, University of Victoria (2024), with Ibrahim Numanagić. Thesis on compilers and secure computational genomics. M.Sc. and B.Sc. in Theoretical Computer Science, University of Sarajevo, with the Golden Badge for highest GPA.
Five years as a lead software, algorithm, and machine learning engineer at Symphony, delivering production systems for US clients including EagleView (3D reconstruction from LIDAR point clouds) and Fathom Health (NLP for medical coding). Also a lecturer in Computational Geometry at the University of Sarajevo.
For project inquiries, a short note describing the problem and timeline is most useful. Typical engagements range from one-week feasibility studies to multi-month implementation projects.