Khine Nyein Aung (Jasper)
Solutions Architect — Vision AI, Video Analytics & Edge AI
Solutions Architect and hands-on engineering lead designing and building Vision AI applications that connect live video, GPU inference, and enterprise workflows. Brings 15+ years of experience in Singapore’s enterprise physical-security sector, including 10+ years focused on video analytics, spanning solution design, proof-of-concept evaluation, systems integration, and field deployment.
Core Competencies & Technical Scope
Vision AI & Multimodal Inference
Engineering practical inference pipelines connecting live camera feeds to lightweight vision models, local Visual Language Models (VLMs), and containerized microservices in Python and PyTorch under strict latency and VRAM constraints.
Video Streaming & Edge AI
Architecting low-latency RTSP and real-time video ingestion, multi-stream GPU optimization, and containerized inference servers on local NVIDIA RTX hardware.
Solutions Architecture & Pre-Sales Discovery
Leading customer discovery, proof-of-concept (POC) latency-vs-accuracy evaluations, edge hardware sizing, and joint partner communication across commercial and industrial deployments.
Enterprise Physical Security Systems
15+ years of experience delivering and integrating enterprise video surveillance, access control, and supporting network infrastructure in Singapore, including hardware I/O integration, fiber backbones, VLAN segmentation, and site risk assessments.
Stakeholder Communication & Technical Presentations
I communicate effectively with stakeholders at all levels, translating user requirements into clear technical proposals and presenting solution approaches, trade-offs, and delivered results in language suited to the audience.
Professional Experience
Trinet Technologies Pte Ltd
Sep 2017 — Present • SingaporeHead of Technical & Lead Solutions Architect — Vision AI & Video Analytics
- Serve as the first line of technical expertise for enterprise customers, regional partners, and internal engineering—leading technical discovery, live demos, architecture reviews, and GPU/camera sizing across ~30 on-premises GPU video analytics sites and a wider edge-analytics camera fleet.
- Architected and shipped Vision Patrol, an end-to-end visual intelligence platform combining sub-second RTSP-to-WebRTC streaming (<400 ms), multi-stream scheduling, containerized Python GPU inference, semantic vector search (BGE embeddings), and NVIDIA NIM VLM serving.
- Benchmarked and deployed real-time detection and multimodal models (YOLOE-26, D-FINE, DAMO-YOLO, DEIMv2, Qwen3.5) on local NVIDIA RTX Pro 2000 Blackwell and RTX 2070 GPUs, balancing accuracy, throughput, and VRAM budgets to cut false-positive operator alerts by 85%+.
- Built a natural-language video search and summarization engine following the NVIDIA VSS pattern with 2x2 / 3x3 multi-camera grid compositing—cutting VLM token use and GPU compute overhead by up to 75%—alongside automated VLM agent evaluation pipelines.
- Designed and commissioned closed-loop Physical AI workflows linking MQTT facial-recognition terminals and alarm I/O relays with real-time GPU visual verification.
- Drive technical alignment and co-selling with global video AI and security partners (IronYun, Univers AI, Convergint) on major enterprise tenders, while recruiting, mentoring, and leading a 17-member engineering department across 10G network infrastructure and site commissioning.
Trinet Technologies Pte Ltd
Sep 2013 — Sep 2017 • SingaporeProduct Manager — Video Surveillance & Security Technology
- Analyzed market trends and customer requirements to source, introduce, and scale a major surveillance brand in Singapore into multi-million-dollar enterprise deployments.
- Partnered with sales and engineering teams to translate complex client requirements into structured technical proposals, live demonstrations, and tailored security architectures.
Trinet Technologies Pte Ltd
Sep 2011 — Sep 2013 • SingaporeApplication Engineer — Video Surveillance, Access Control & Intrusion Detection
- Designed, configured, and deployed custom IP/analog video surveillance, access control, and intrusion detection systems tailored to commercial and industrial site requirements.
- Supported pre-sales cycles with technical presentations, proof-of-concept demonstrations, and tender proposal engineering.
Selected Engineering Architectures
Vision Patrol
Lead ArchitectAutonomous visual-intelligence platform for virtual guard tours, combining low-latency multi-camera streaming, local VLM scene reasoning, and automated incident reporting.
Team Reports
Lead ArchitectEnterprise field-service platform connecting multi-team job dispatch with automated compliance reports, mobile offline sync, and edge API microservices.
RuneCam
CreatorCross-platform mobile inspection utility for site engineers to capture commissioning photos, apply vector overlay annotations, record voice notes, and export standardized PDF reports on-site.
Education & Credentials
Bachelor of Engineering in Information Technology
Yangon Technological University
AWS Certified Solutions Architect – Associate
Amazon Web Services (SAA-C02)
Previously Certified
Cisco CCNP — Routing & Switching
Cisco Networking Curriculum
Advanced enterprise routing, switching, and network segmentation