AI Video Analytics
End-to-end detection pipelines on NVIDIA DeepStream — face recognition, ANPR, PPE, fire/smoke, crowd counting — tuned for real camera feeds, not benchmark clips.
DeepStreamYOLOTensorRTMulti-streamRe-ID
Open for AI / CV & R&D roles
Ahmad Farid Muharram
AI & Computer Vision Developer
I build production AI video analytics — from the GPU pipeline to the dashboard the client actually uses.
statusopen to work
roleR&D Supervisor
focusDeepStream · ANPR · FR
stackPython · YOLO · TensorRT
local--:--:-- WIB
reply< 24 jam
2+
Years leading R&D
Supervisor level
6
Shipped AI modules
FR · ANPR · PPE · Fire
National
ETLE deployment scale
Korlantas Polri
24/7
Uptime target per site
Edge + on-prem GPU
01 — About
Engineer first, supervisor second.
R&D Supervisor architecting multi-tenant AI video analytics platforms (DeepStream/YOLO) for face recognition, ANPR, and real-time detection at scale. I pair hands-on engineering — edge AI, GPU optimisation, custom model training — with team leadership, project management, and direct client engagement.
I lead the R&D function at Office and Security Solution, where my team designs and ships an AI video analytics platform that runs on real camera infrastructure across Indonesia — police, city government, industrial sites.
Most of my work sits at the messy intersection of model, GPU, and network: squeezing a detector into a Jetson budget, keeping 40+ RTSP streams alive on one DeepStream pipeline, reconciling what a demo promises with what a site survey allows.
The other half is people and paperwork — scoping with clients, sizing hardware, running procurement, and translating "we want smart CCTV" into something an engineer can actually build.
02 — Services
What I can help with
Four things I get asked for most — usually two or three of them on the same project.
Platform & Integration
Multi-tenant dashboards, event APIs, message brokers, and the plumbing that turns raw inference into something operations teams can act on.
PythonREST APIKafkaMySQLDashboards
Edge & Infrastructure
GPU sizing, on-prem and edge deployment, VPN meshes between sites, and camera network design that survives a real install day.
JetsonMikroTikWireGuardTrueNASSite survey
Technical Leadership
Running R&D roadmaps and annual government projects: scoping, budgeting, vendor coordination, commissioning, and client-facing technical defence.
Project mgmtINAPROCTeam mentoringClient demo
03 — Selected work
Systems that run on real cameras.
Production deployments, not demos. Click a card for the scope, the constraints, and what I actually shipped.
AI PlatformLive demo01 / 2024 — nowProduction
OSS Vision / SWard
Architect & Lead · Office and Security Solution
Multi-tenant AI video analytics platform on DeepStream/YOLO. Full pipeline from edge capture to dashboard, deployed across Indonesian sites with analytics modules clients mix per deployment.
Government
02 / 2024 — nowOngoing
ETLE Korlantas Polri
Annual Project Manager · Korps Lalu Lintas Polri
Electronic Traffic Law Enforcement camera deployment for the National Traffic Police. Site surveys, installation, integration, and commissioning across multiple provinces.
Smart City
03 / 2024Delivered
SIGAP / Seribu Wajah
R&D Lead · Diskominfo Bandar Lampung
Real-time AI CCTV analytics for Bandar Lampung city government — live face recognition and detection running across public camera feeds.
Commercial
04 / 2024 — 2025Production
GNS — Safety & Compliance Suite
Solution Design · GNS
Commercial deployment of OSS Vision analytics: fire/smoke detection, face recognition, PPE compliance, and crowd monitoring, with hardware sizing and pricing.
04 — Career
From splicing fiber to leading R&D
Ten years of moving up the stack — cables, then networks, then code, then the team that writes it.
- Annual Project Manager for the ETLE Korlantas Polri national deployment
- Lead and mentor a cross-functional R&D team — engineers, researchers, annotators
- Architect a multi-tenant AI video analytics platform on DeepStream/YOLO
- Manage GPU, hardware, and budget resources against timelines and technical risk
- Report to senior management and lead client technical presentations and Q&A
DeepStreamYOLOTensorRTPythonGPUProject management
05 — Toolkit
Things I reach for
Grouped by what they're for rather than by logo — everything here has shipped in something real.
NVIDIA DeepStream
YOLO (v5–v11)
TensorRT
OpenCV
PyTorch
Face Recognition
ANPR / OCR
Model Training
Dataset & Annotation
Credentials
Where the hands-on skills came from
Mostly learned on deployment days rather than in a classroom — listed here with the context that earned them.
NVIDIA DeepStream & TensorRT
Self-directed / applied in production
2024 — now · Multi-stream pipelines, GPU batching, model deploymentMikroTik Network Fundamentals
Applied — Fajar Gelora Inti
2022 — 2023 · Routing, VPN, and multi-site camera networksFiber Optic Splicing & Termination
Biznet Networks
2016 · FTTH installation and troubleshooting06 — Contact
Punya project atau posisi yang cocok?
Cerita saja soal kameranya, lokasinya, dan apa yang mau dideteksi — sisanya kita bedah bareng.
GitHub @faridnetzzLinkedIn Ahmad Farid MuharramEmail [email protected]WhatsApp Chat langsung
Location Jawa Barat, Indonesia