Electrical Fault Detection AI Module
Краткое
Freelancer Client is hiring: Electrical Fault Detection AI Module.
Location: Remote
I need a robust ML module that monitors the incoming data patterns and identifies failure patterns in advance. The incoming data is related to power supply parmaters like voltage, current etc. The core of the job is to take my historical performance logs and a live stream of monitoring data, train an accurate model, and expose predictions through a clean, well-documented interface that my team can plug straight into our existing control software.
What you'll do:
• Data handling: build an ingestion pipeline that pulls real-time feeds (Kafka or MQTT are fine) alongside batch uploads of past performance files, then stores everything in a format that supports fast feature extraction.
• Model development: use Python with TensorFlow, PyTorch, or an equivalent deep-learning framework to train a classifier/anomaly detector that flags incipient and critical grid faults. Please include explainability techniques so our operators can trust the alerts.
• Deployment: wrap the model in a lightweight REST or gRPC service, complete with health checks and graceful fail-over logic suitable for on-prem or cloud (Docker/Kubernetes).
• Testing & metrics: supply unit tests, performance benchmarks, and a validation report showing precision/recall on unseen data.
• Documentation: concise setup guide and API reference.
• 1. End-to-end pipeline processes both historical and live data with <5 s latency on real-time streams.
• 2. Model meets or exceeds 95 % fault-detection recall on our validation set while keeping false positives below 3 %.
• 3. Containerised service starts in under 30 s and passes all included tests on a clean machine.
Skills: Python, Machine Learning (ML), MQTT, Deep Learning, Anomaly Detection, REST API, Model Deployment
Budget: $12500–$37500 USD
Source: Freelancer Client via Remote / Online. Apply on the source website.
Оригинал
I need a robust ML module that monitors the incoming data patterns and identifies failure patterns in advance. The incoming data is related to power supply parmaters like voltage, current etc. The core of the job is to take my historical performance logs and a live stream of monitoring data, train an accurate model, and expose predictions through a clean, well-documented interface that my team can plug straight into our existing control software.
Here is how I picture the workflow:
• Data handling: build an ingestion pipeline that pulls real-time feeds (Kafka or MQTT are fine) alongside batch uploads of past performance files, then stores everything in a format that supports fast feature extraction.
• Model development: use Python with TensorFlow, PyTorch, or an equivalent deep-learning framework to train a classifier/anomaly detector that flags incipient and critical grid faults. Please include explainability techniques so our operators can trust the alerts.
• Deployment: wrap the model in a lightweight REST or gRPC service, complete with health checks and graceful fail-over logic suitable for on-prem or cloud (Docker/Kubernetes).
• Testing & metrics: supply unit tests, performance benchmarks, and a validation report showing precision/recall on unseen data.
• Documentation: concise setup guide and API reference.
Acceptance criteria
1. End-to-end pipeline processes both historical and live data with <5 s latency on real-time streams.
2. Model meets or exceeds 95 % fault-detection recall on our validation set while keeping false positives below 3 %.
3. Containerised service starts in under 30 s and passes all included tests on a clean machine.
If this matches your expertise, let’s talk timeline and milestones so we can move quickly from prototype to production.
Локация & Details
Перейти к источнику →About this listing
This remote opportunity was imported from Freelancer and is shown here for discovery. To apply, follow the link to the original posting.