What Our Clients Have to Say
Real experiences from organisations that have partnered with us on their computer vision and AI projects.
Back to HomeClient Reviews
Nurul Zahirah
Operations Manager, Shah Alam
We brought pulsarars in to build a defect detection system for our packaging line. The team spent genuine time understanding our production environment before proposing a solution. The system has been running for three months now, and the detection accuracy matches what they projected during scoping. Communication was clear from start to finish.
28 January 2026
Chong Wei Lim
IT Director, Penang
The video analytics platform they built integrates well with our existing CCTV setup. Setup took about eight weeks, slightly longer than expected due to some camera compatibility issues, but pulsarars handled it well and kept us informed. The people-counting feature has been useful for our retail space planning decisions.
21 January 2026
Priya Menon
Research Lead, Kuala Lumpur
We used their data preparation service to structure a large set of microscopy images for a classification project. Their annotation guidelines were thorough, and the QA process caught inconsistencies that would have caused headaches downstream. The versioned datasets they delivered made our subsequent model training much smoother.
14 January 2026
Amir Kamal
Facilities Manager, Johor Bahru
Straightforward team to work with. They implemented a video analytics system across three of our warehouse locations. What I appreciated most was that they didn't oversell — they were upfront about what the technology could and couldn't do for our use case, and the final system performed exactly as scoped.
5 February 2026
Sarah Tan
CTO, Cyberjaya
We engaged pulsarars for an image classification project in our quality control department. The model performs well for most product variants, though we are still working together to refine accuracy for a few edge cases. Their willingness to continue iterating post-delivery has been reassuring — it feels like an ongoing partnership rather than a one-off hand-off.
10 February 2026
Rizal Hashim
Logistics Coordinator, Selangor
The data preparation service helped us turn a messy collection of warehouse images into a clean, annotated dataset. What impressed me was the quality assurance process — every batch went through checks, and they flagged ambiguous cases for our review rather than guessing. Now we have a solid foundation for building our own detection models.
30 January 2026
Success Stories
A closer look at three engagements that illustrate how our services translate into tangible outcomes.
Manufacturing Quality Gaps
A Selangor-based electronics manufacturer was losing approximately 4% of production to undetected surface defects. Manual visual inspection was slow and inconsistent across shifts, leading to customer complaints and returned batches.
Custom Vision System
We developed a computer vision system trained on 12,000 annotated images of both defective and acceptable units. The model was optimised for the client's existing inspection station hardware, delivering sub-200ms inference per image.
Measurable Improvement
Defect escape rate dropped from 4% to under 0.8% within the first month of deployment. The system processes units 3× faster than manual inspection, and the client has since extended the system to a second production line.
Timeline: 11 weeks
Retail Foot Traffic Blind Spots
A mid-size retail chain in KL had CCTV cameras across eight outlets but no way to extract useful data from them. Store managers relied on subjective estimates for foot traffic and customer dwell patterns.
Video Analytics Deployment
We integrated our video analytics platform with their existing camera infrastructure, configuring people counting, zone-based dwell-time tracking, and a centralised dashboard accessible by area managers.
Data-Driven Decisions
Store managers now make staffing and layout decisions based on actual traffic data. The client reported a 12% improvement in staff scheduling efficiency within the first quarter, and the system is now being rolled out to additional outlets.
Timeline: 8 weeks
Unstructured Image Archive
An agricultural research team had accumulated over 50,000 crop images across multiple growing seasons, but the data was unlabelled and stored without consistent structure — making it unusable for training disease detection models.
Structured Data Pipeline
We created detailed annotation guidelines in consultation with the research team's plant pathologists, then annotated and versioned the full dataset with bounding boxes and classification labels across seven disease categories.
Research-Ready Dataset
The team received a clean, versioned dataset with inter-annotator agreement above 93%. They used it to train a baseline detection model that is now part of an ongoing field trial. The client has since engaged us on a retainer basis for new data each season.
Timeline: 5 weeks
By the Numbers
Years of Operation
Projects Delivered
Average Client Rating
Repeat Engagement Rate
Get in Touch
Have questions about how we can help? Reach out through any of these channels.
Monday – Friday: 9:00 AM – 6:00 PM · Saturday: 10:00 AM – 2:00 PM
Join the Organisations That Trust pulsarars
Start a conversation about your computer vision needs. We will listen first, then suggest the best path forward.
Start a Conversation