Participant Record · July 2025
Feedback collected via written questionnaire three weeks after course completion. Case studies describe engagement processes and outputs, not claimed outcomes.
Back to Home240+
Course participants since 2020
4.6 / 5
Average post-course rating
19
Consulting engagements
6 yrs
Operating in Penang
Ruzaini Hassan
Product Manager · Kuala Lumpur
The short course clarified several things I had misunderstood about how language models produce their outputs. The bibliography they handed out on day one was useful — I went back to three of the papers after the course. The final session on what we had not covered was an unusual choice but it made sense. I came in with a loose understanding and left with a more careful one.
Introduction to AI Systems · June 2025
Siew Cheng
ML Engineer · Penang
The evaluation course was the most useful nine weeks I have spent on professional development. Building the evaluation suite for my own model was demanding but that is the point. The statistical significance section was tightly taught and the red-team session raised failure modes I had not considered. The written report required clear thinking about weaknesses in my own work, which is not comfortable but is valuable.
Model Evaluation and Reliability · May 2025
Nurul Ain
Business Analyst · George Town
I work in a team that reviews AI-generated reports and I was not sure what I was actually assessing. The short course gave me a framework for thinking about where errors come from. I do wish the section on evaluation metrics had more time. The evening format worked well given my schedule. The fact that they list what they will not teach in advance meant I came in with accurate expectations.
Introduction to AI Systems · June 2025
Tan Wei Liang
Technical Writer · Penang
The sourcing practice was what stood out. I write technical documentation for AI systems and having course materials that demonstrate how to attribute claims properly was directly applicable. The section on why model outputs vary was clearly explained without oversimplifying. A solid four weeks.
Introduction to AI Systems · July 2025
Ahmad Irfan
Data Scientist · Shah Alam
The ablation design and calibration sessions in the evaluation course were the strongest parts for me. I found the production monitoring section slightly rushed relative to how important it is, but there is only so much you can fit in nine weeks. What I produced for the assessment — a full evaluation suite for a model I had been using at work — was genuinely useful. Four stars for depth, would have liked more time on the monitoring component.
Model Evaluation and Reliability · April 2025
Lim Mei Ling
Operations Lead · Butterworth
I sent two of my team members to the short course before we started working with a vendor on an AI integration. The course gave them a vocabulary for asking better questions during the scoping meetings. The vendor was surprised by how specific the questions were. Worth the two weeks of evening sessions.
Introduction to AI Systems (Team Enrolment) · June 2025
The team had deployed a document classification model into their operations intake process without a systematic way to monitor whether it was declining in accuracy over time. Errors were caught manually weeks after they occurred.
Discovery workshop identified that no structured evaluation baseline had been established at deployment. Cogniva produced a written review, a monitoring workflow recommendation, trained four engineers and two product managers over three days, and provided a set of documentation templates for logging model behaviour.
At the four-month review, the team had implemented weekly accuracy spot-checks using a documented holdout set and had a written incident log covering two detected issues. The process was documented and assigned to a named owner.
A product team was incorporating LLM-generated content into a client-facing report tool. No internal review practice existed. Engineers and product managers had different understandings of what the model could and could not produce reliably.
Discovery session with engineering and product leads identified gaps in shared understanding. Cogniva produced a written assessment of the current review process, a recommended evaluation workflow based on published frameworks, trained six staff over three days, and delivered a documentation template set covering output sampling and review logging.
At the four-month review, the team had a shared written definition of output quality criteria, a sampling schedule in place, and a backlog of documented edge cases. The product manager reported that disagreements about model reliability were now resolved by reference to written criteria rather than by escalation.
Telephone
+60 4 3915 6280Address
33 Jalan Sungai Pinang
10150 George Town
Pulau Pinang, Malaysia
Office Hours
Monday – Friday: 9:00 – 18:00
Saturday: 9:00 – 13:00
Penang ICT Council — 2023 Education Partner
Recognised for structured AI education provision in Penang's technology and manufacturing sectors.
MDEC Registered Training Provider (since 2021)
Listed on the Malaysia Digital Economy Corporation training provider registry.
HRD Corp Claimable
All three programmes are eligible for Human Resources Development Corporation claims for Malaysian-registered employers.
Send a message with your name, role, and which programme you are considering. We will confirm availability and respond within two working days.
Send an Enquiry