Medical student. Open source contributor. Specification driven researcher.
Second year medical student at Menoufia National University, Egypt. I contribute to an open source clinical NLP library, reproduce published statistics papers and correspond with their authors to verify correctness, and design machine learning research projects that are implemented and tested through directed AI coding agents.
Eight pull requests submitted, merged to maziyarpanahi/openmed, spanning data quality (golden fixtures, patient record span filtering), reproducibility and integrity tooling (hash verification, skill bundle signature checks), model infrastructure (teacher ensemble registry), documentation (an executed notebook gallery with CI checks), privacy (notebook cell redaction), and clinical interoperability (a FHIR DiagnosticReport exporter).
Reproduction of Masselot et al., CIRLS GLM Independently reproduced the results of this peer reviewed climate and health statistics paper without institutional compute access. Lead author Pierre Masselot and co-author Antonio Gasparrini, London School of Hygiene and Tropical Medicine, confirmed by email that the reproduction was correct.
Streaming Time Series Anomaly Detection Methodology designed and implementation directed against the TSB AD U benchmark (350 series). Result: 70 percent win rate against a persistence baseline on VUS ROC, 51 percent on VUS PR. Corresponded with the author of a published bearing degradation model to validate a proposed synthetic fault onset benchmark.
- cw-node-research, connection weighted node architecture. Results published at mohamedhossammohamed.github.io/cw-node-research
- open-phase-ensemble
- overfit-init-transfer
- agent-parallel-merge
- FORGE
- m4-prefill-engine
Chimera Agent Medical AI Grand Challenge team project. Built exploratory case data analytics for a prostate cancer risk modeling task. The team did not submit to the challenge.
Languages: Python, working with PyTorch, MLX, FastAPI Tools: Git and GitHub, AI coding agent orchestration (specification driven development) Spoken: Arabic (native), English (fluent)