Available for Enterprise Deployments

Building Vision AI pipelines for complex environments.

Vision AI Team Lead with 5+ years of experience architecting end-to-end edge computer vision systems and applied LLM verification layers. Select an industry below to test live production sandbox simulations.

Automotive & Smart City

Dashcam processing, plant safety compliance, and edge LPR deployments.

Explore Sandbox

Oil & Gas / HSE

LLM-assisted cleanliness monitoring and strict PPE compliance pipelines.

Explore Sandbox

Infrastructure

Drone telemetry, road defect segmentation, and remote security patrols.

Explore Sandbox

Automotive & Smart City

Edge deployments for plant mobility and urban analytics.

Custom YOLOv8 models deployed on city CCTV edge nodes for automatic number plate recognition (ANPR), cross-referenced in real time against a blacklist database to flag wanted or stolen vehicles.

PLATE: VFE9920 — CLEAR ✓
Inference: 12ms

Vision AI monitoring for safety compliance within the Perodua Smart Mobility Plant, ensuring operational boundaries and machine interaction safety protocols are followed.

PERSON_SAFE (0.92)

Oil & Gas / HSE

High-precision monitoring and LLM verification layers.

LIQUID_SPILL (0.88)
LLaMA Verification
"Context confirms reflection on metallic surface, not a chemical spill. Suppressing false positive."

Deployed at Petronas Research to monitor lab cleanliness. Standard CV pipelines often flag light reflections as spills. I integrated an LLM reasoning layer (Qwen/LLaMA) to verify context and slash false positives.

NO_HELMET (0.95)
SAFETY_VEST (0.91)

Object detection and segmentation models (YOLO, PyTorch) customized for HSE and PPE compliance monitoring to ensure site safety standards are strictly met.

Infrastructure

Drone telematics and remote inspections.

Dashcam and drone-based road operations deployed in Malaysia, Singapore, and Japan. Utilizes object segmentation to precisely map pothole surface areas and structural cracks.

POTHOLE_DEFECT

Wide-area aerial security patrols utilizing PyTorch detection pipelines. Designed to identify unauthorized access or structural anomalies from high-altitude moving camera feeds.

INTRUDER (0.87)
STATUS: CLEAR