About Kaedris
Kaedris is led by Sahel Iqbal, a machine learning researcher and engineer currently working as a postdoctoral researcher at the University of Oxford.
Sahel completed his PhD in machine learning at Aalto University in Finland, and started Kaedris to bring the benefits of cutting-edge AI into dependable workflows with clear controls, clear ownership, and clear numbers.
Founder-led
Kaedris is intentionally small and technical. The work is shaped by research discipline: define the business problem, understand the baseline, test a narrow intervention, and scale only when the result is useful.
Founder
Machine learning researcher and engineer focused on moving AI from impressive demos into reliable operating systems for real teams.
Current role
Postdoctoral researcher, University of Oxford
Doctoral training
PhD in machine learning, Aalto University, Finland
Kaedris focus
Cutting-edge AI applied to measurable SME workflows
Why we exist
Most teams have already seen what a modern AI model can do. That does not automatically make it ready for customer messages, invoices, collections, reporting, or operational decisions.
Business use needs the parts around the model: the right data, the right permissions, human review where it matters, error handling, logs, escalation paths, and integration with the software your team already trusts.
We bring state-of-the-art AI knowledge into practical business systems, then keep the scope narrow enough that the result can be measured.
The AI value gap
This is the problem Kaedris was built to solve. AI does not pay off because it is fashionable. It pays off when it is attached to a real workflow and a real business metric.
Evidence
Scaling gap
Only 39% report EBIT impact at scale.
AI usage is widespread, but the financial impact is still uneven. The difference is usually not access to a model. It is whether the model is embedded in a workflow that changes how work gets done.
Our motivation
01
A personal chatbot can help someone write, summarize, or brainstorm. A business workflow has to be reliable, secure, observable, and connected to the software the team already uses. It needs model judgment, integrations, guardrails, review loops, and clear ownership when something needs human attention.
02
Many SMEs in India and the UAE cannot justify large transformation teams or global consulting fees. They still deserve deep machine learning knowledge applied to local tools, local budgets, and workflows where the business case can be proven.
How we work
01
We work backward from the business owner, the manager, and the employee doing the work every day. If a workflow does not make their life easier or their numbers better, it is not worth building.
02
AI projects should earn the right to expand. We define the baseline, choose the KPI, and prove whether the pilot improves revenue, response time, cash collection, throughput, or manual effort.
03
Useful AI has to meet teams inside their current operating system: WhatsApp, Excel, Tally, email, CRMs, shared drives, and approval habits that already run the business.
04
Our goal is not to replace good people with brittle automation. It is to remove repetitive coordination, data entry, and follow-up so teams can spend more time on judgment, relationships, and creative problem solving.
We will help you find where AI can actually improve revenue, margin, response time, or manual workload before you commit to anything bigger.
