About the practice
Operational experience.
Research-led direction.
MEKreflect is the consultancy and applied AI research practice Haris Mekic is building around a simple conviction. Technology becomes useful when it understands the work and helps people become more capable.
Haris Mekic
The person
behind the direction.
My background is in institutional operations and project delivery. Working with experts, stakeholders, budgets and deadlines taught me that the difficult part is often not the plan. It is the context people need to carry it out.
I have always loved computers and being part of that world. AI shortened the distance between an idea and something I could build, open and test. I bring the problem, ask questions, shape the context and direct implementation with AI tools. Then I examine what the system actually did.
My political-science background also shapes the questions I explore. How does information influence decisions? What changes when an assumption changes? Can we repeat the method and understand why an outcome moved?
I started MEKreflect to give that work a shared direction. I want to build with people who care about practical results and learning, and help teams carry their own knowledge into better working systems.
A digital office.
A human purpose.
The ambition is people and specialist AI workers collaborating through shared context and clear responsibilities.
This is a founder-led practice in development. The next stage is scoped collaboration and supervised use with teams, measured through the quality of accepted work and the people able to continue it.
How the method worksHow I turn a problem
into working software.
I start with the work and the people responsible for it. I turn their requirements into shared context, role-based procedures and tools that can be tested. Then I inspect what the workers passed to each other and what the final action depends on.
My development work uses AI-assisted Python and TypeScript, model APIs, structured data and human review. Inspected implementations include specialist workers sharing a task context, user and organisation permissions around tool calls, and database queries kept within an organisation’s data boundary.
Supabase, Vercel and Cloudflare are part of that development history. Using a platform is one part of the work; understanding the source records, access rules, failure paths and maintenance requirements is what makes the implementation useful.
That is the practical expertise I bring into research, workflow design and team learning.