AI Engineer
Send a job offer directly to this candidate
AI Engineer with 3+ years of experience specializing in Computer Vision, VLM, and LLM-based systems. Proven track record building and deploying end-to-end real-time video analytics platforms at enterprise scale using NVIDIA DeepStream, TensorRT, YOLO, and GPU acceleration. Published researcher (Google Scholar, ACM) with strong background in embedded systems, IoT, and system-level architecture.
Passionate about building practical, high-performance AI systems that solve real-world operational problems.
https://scholar.google.com/citations?user=ri0sWlAAAAAJ
Architected and built solo PT United Tractors' (Astra Group) first CV deployment — a 70-camera real-time AI platform for smart warehouse CCTV at ~20
FPS per stream on 2x RTX A4000 GPUs. Designed 15+ AI models (80-90% mAP@0.5) including PPE, fire/smoke, activity recognition, ANPR. Trained all models on 172K+ annotated images. Built Ciptakom (Ciputra Group) 10-channel video analytics with vehicle detection and ANPR. Conceived CV-based fraud detection system for Warjon Restaurant. R&D: VLM-based CCTV narration, Grounding DINO auto-annotator, CogVideoX synthetic data generator.
Main engineer for SCORE Hub — modular industrial IoT middleware unifying Modbus, OPC-UA, EtherCAT, and REST. Built Flask backend, protocol modules,
RBAC auth, Docker microservices, and full client documentation.
Owned full-stack embedded development for smart air cleaner — ESP32 firmware, TFT touch UI, PM2.5 sensing, Modbus RTU fan control, Wi-Fi with
WebSocket telemetry and OTA updates. Delivered production-grade prototype.
Led bionic hand prosthesis research using EMG signal processing and ML classification. Published at ACM SIET 2023 (Bali) and ICOMIT. Deployed optimized models on ESP32/ARM.
Led IoT-based Autonomous Tram Mover prototype (G20 initiative) — CBTC compute engine, MQTT telemetry, GPS safety-distance logic, and operator control center UI.
