Founder
Builds systems whose output has to hold up when somebody disagrees with it.
Ramandeep Singh Aulakh founded Digiaeon Services Pvt Ltd to do one kind of work well — design, build and operate software that other people depend on daily. He is based in Chandigarh and works directly on engagements rather than above them: architecture, evaluation design, and the scoping call that decides whether a project has a real first cut or only a roadmap.
Education technology came first, and still does. He joined TCY Learning Solutions in February 2015 as Deputy Manager, Marketing, and left in 2018 to found Exoways Web Technologies. Exoways operates FourModules, a test-preparation platform for CD-IELTS, PTE, CELPIP and CAEL — candidates are scored by it, and dispute the score when the official one differs.
That is the through-line. In test preparation the output is graded, contested and has to be defensible to the person on the receiving end. Applied AI is the same discipline under a different name: a score or an answer is worth nothing unless you can show how it was produced, measure whether it got better, and explain the failure when it did not. Digiaeon is built around that constraint rather than around the model of the month.
February 2015 — 2018
TCY Learning Solutions Pvt Ltd
Three years inside an established test-preparation business — where demand is seasonal, the product is a score, and the distance between a marketing claim and a candidate’s actual result is very short.
2018
Exoways Web Technologies Pvt. Ltd.
Left TCY to start Exoways — a product company rather than a services shop, where the team owned what it shipped, including the parts that broke.
Ongoing
FourModules, an Exoways platform
An online test-preparation platform for CD-IELTS, PTE, CELPIP and CAEL. Four exams, four rubrics, four sets of edge cases — and a scoring pipeline that has to stay consistent while content, formats and candidate volume keep moving.
Now
Digiaeon Services Pvt Ltd
Applied AI, data platforms, custom software and the infrastructure underneath them, for companies that need a system in production rather than a pilot that presents well.
Positions
Anything that reaches real users needs an owner, a rollback path, a budget ceiling and somebody who gets paged. If those four are missing, the demo was the whole project — and the cost of finding that out arrives later, with interest.
Most AI work goes wrong because nobody defined success precisely enough to argue about it. Build the eval set first, from real traffic, including the cases you would rather not look at. After that, every change is a decision instead of an opinion.
Postgres, a queue, typed contracts and a deploy pipeline that runs on every merge. Novelty belongs in the part of the system that faces the customer, not in the part that has to be awake at 3am.
Stating the limits early costs one uncomfortable conversation. Discovering them in production costs trust, and trust is the only thing a small firm actually sells.
Direct line
Early conversations are with him directly — bring the problem, not a requirements document.