Engineer & Researcher
BS Computer Engineering (COMSATS University). MS Artificial Intelligence Engineering. Research in efficient vision model deployment on constrained embedded hardware and measurement methodology for compression decisions.
Measure It. Don't Assume It.
My research concerns how decisions about vision model compression are made, and what information is actually needed to make those decisions reliably on the hardware that will run the result. It is grounded in measurement: running configurations on the target device under controlled conditions rather than relying solely on structural estimates.
The work builds a methodology for evaluating compression decisions against what is actually observed on constrained hardware. Details of the findings are being prepared for publication. More on the Research page.
Through Pesgit Engineering, I apply this methodology to real engineering problems, building vision systems on embedded hardware with performance claims justified by measurements from the target device.
“You cannot optimise what you have not measured, and you cannot trust a measurement you cannot reproduce.”
Engineering From First Principles
Formal training in both the hardware and the AI, combined into research that lives at their intersection.
BS Computer Engineering
COMSATS University Islamabad. Foundation in computer architecture, digital signal processing, embedded systems, and software engineering, the hardware and systems layer beneath every deployment decision I make.
MS Artificial Intelligence Engineering
Advanced study in machine learning, computer vision, neural network architecture, and real-time inference optimisation. Research into efficient deployment on constrained hardware, with a specific focus on the measurement methodology needed to make compression decisions trustworthy.
Research Output
The MS research produced technical output at the intersection of model efficiency and measurement methodology for efficient vision model deployment on constrained hardware. Details available on the Research page.
My Operating Principles
Architectural Rigour
I approach problems with structural discipline. The underlying data model and architecture decide whether the result holds; I design from the structure up.
Clarity Over Complexity
I impose order on complexity. My primary objective is to transform fragmented data and siloed teams into structured, reliable decision systems.
Execution That Ships
Effective systems demand follow-through. Every strategy I build, across both my ventures, is anchored to something real, working, and measurable.
Pragmatic Design
I design frameworks, whether they are AI perception modules or corporate KPI hierarchies, that survive real-world use. Usability and reliability over theoretical perfection.
Applying the Research in Practice
The research is not separate from the engineering; it informs what gets built and how it gets tested.
Through Pesgit Engineering (Rawalpindi, Pakistan), I run applied R&D in edge AI and computer vision, building vision systems on constrained, offline-capable hardware and validating every performance claim with measurements from the target device rather than analytic estimates.
Through Femtus Solutions (London, UK), I apply the same rigour to data and reporting, building management reporting systems, performance analysis frameworks, project controls, and reporting automation that give leadership teams clear, reliable visibility.