Muhammad Asim

Muhammad Asim is the founder and CEO of Devsort, a clinical-grade data science and software engineering company based in Islamabad, Pakistan. With more than five years working on medtech data science and regulated software, he founded Devsort to focus exclusively on technically complex, high-stakes engineering work — the kind where the cost of getting it wrong is real and the credibility of getting it right is verifiable.

The engagement that defines Devsort’s practice is a 2.5-year programme with PKG Health (now part of Empatica), a company whose algorithms for monitoring movement disorders in Parkinson’s disease patients are cleared by the FDA under the 510(k) pathway, CE marked, and TGA approved. When Asim’s team took on the engagement, the algorithms were frozen — implemented across four legacy programming languages, undocumented, and effectively unmaintainable under the regulatory constraints that governed any change to a certified system. The brief was to refactor all six algorithms without invalidating the existing certification. Over the engagement, the algorithms were refactored and validated in Python, extended to five wearable device platforms, and a subset deployed to embedded C on the Empatica device hardware, reducing tremor algorithm processing time from five minutes to thirty seconds.

The technical and regulatory discipline from that engagement informs every project Devsort takes on. Working under FDA 510(k), CE, and TGA frameworks simultaneously over 2.5 years produces a specific kind of engineering practice — one that treats documentation as a first-class deliverable, validates outputs against reference implementations rather than intuition, and understands that regulatory constraints are the specification, not an obstacle. Asim applies that discipline whether the project is a clinical algorithm, a wearable preprocessing pipeline, embedded firmware, or a software platform for a healthcare company.

He writes on the Devsort blog on topics including clinical algorithm validation, wearable data science, and FDA-compliant software development. All technical content on this site is written from first-hand experience — not from secondary sources.