Python Developer – Video Analytics Pilot (Manufacturing Floor)
We are looking for an experienced Python engineer to design and build a lightweight, cloud-assisted video analytics pilot for a manufacturing floor using our existing camera infrastructure.
The objective is to monitor worker compliance during shift hours using efficient image sampling and Vision AI, avoiding heavy continuous video processing.
We currently operate a Dahua NVR608RH-32-XI system with built-in Face Recognition and want to leverage existing capabilities rather than build custom computer vision pipelines.
We are looking for someone who can propose and implement a pragmatic, maintainable architecture within a small pilot scope.
Project Scope
1. Data Ingestion (Edge Script)
Develop a lightweight Python service running on a local PC.
Connect to Dahua NVR via CGI/HTTP API.
Capture periodic snapshots (~1 image every 10 seconds) across up to 32 camera channels.
Listen to the face recognition event stream exposed by the NVR.
2. Shift Management Logic
Integrate a Google Sheet used as a simple shift database.
The system should only analyze images when assigned workers are on shift for a given camera/zone.
3. Face Recognition Sync
Pull recognition metadata from the NVR.
Map detected faces to worker identities maintained by HR.
Maintain a simple repository where identities can be updated without code changes.
4. Cloud Vision AI Analysis
Send sampled images to a Vision AI API (e.g., OpenAI, Gemini, Anthropic).
Classify worker state, for example:
Actively working
Loitering / absent from station
Using a mobile phone
The above approach reflects our current hypothesis. Alternative implementations achieving the same objective efficiently are welcome.
5. Alerting & Logging
When a violation is detected:
Log the event with timestamp and snapshot reference (database or Google Sheet).
Generate a secure link to the image snapshot.
Send immediate alerts via Email or WhatsApp API (Twilio / Meta) identifying the worker by name.
Support automated monthly incident summaries per worker.
Deliverables
Dockerized Python service deployable on a local machine.
Google Sheets integration for shift management and logging.
Email / WhatsApp alert integration.
Clear documentation covering:
Adding new cameras
Updating shift configuration
Running and maintaining the system.
(Authentication via HTTP Basic/Digest.)
Ideal Experience
Strong Python backend or systems scripting experience
Integration with IP cameras or edge devices (Dahua experience is a plus)
Experience integrating Vision LLM / AI APIs
Dockerized deployments
Practical system design for small production pilots
Engagement Details
The pilot is scoped at approximately 40–60 hours; please include your own estimate.
Please include:
Your estimated effort breakdown
Assumptions or risks you see
Suggested improvements (if any)
This engagement is an initial pilot. If successful, we expect follow-on work including system hardening, expanded analytics, additional integrations, and production deployment.
You will have direct access to technical stakeholders and fast feedback cycles.
Contract duration of 1 to 3 months.
Mandatory skills: Computer Vision, Python