Driver Behavior Scoring ML Model Development

Job ID: 39529853

Budget: $250 – $750 USD

Develop Machine Learning Model for Driver Behavior Scoring Using Telematics + Camera Events

Description:
We are Prometheus, a leader in fleet telematics and IoT, and we're seeking an experienced Machine Learning Engineer or Data Scientist to build a model that analyzes driving behavior using GPS and camera-triggered event data.

Objective:
Develop a scoring model that processes data collected every 5 seconds from our telematics and dash camera systems to generate a real-time or batch-based Driver Safety Score.

We Will Provide:
A sample 10–15 minute anonymized dataset in CSV format
Raw GPS/telemetry data:

Timestamp

Latitude / Longitude

Speed

Altitude

Cell ID

Other GPS metrics

Camera-triggered event flags:

Speeding on Posted Limit

Following Too Close

Collision Warnings

Distracted Driver

Eyes Closed

Phone Use

Seatbelt Violations

(More to be added)

Your Scope of Work:
1. Feature Engineering

Create derived metrics (e.g., speed delta, event frequency, acceleration)

Align GPS and event data into a consistent time-series format

2. Baseline Model Development

Use models like Logistic Regression, XGBoost, or Random Forest

Predict risk level or generate a continuous risk score

Output feature importance for interpretability

3. Score Calibration

Normalize output to a 0–100 Driver Safety Score

Define scoring zones: Green (Safe), Yellow (Caution), Red (Risky)

Allow for configurable weights per behavior/event

4. Final Deliverables

Python code or Jupyter Notebook with full model pipeline

Documentation of feature logic and scoring model

Input/output CSV templates

Optional: REST API wrapper for inference

Ideal Candidate:
Experience with time-series or telematics data

Strong in Python, pandas, scikit-learn, XGBoost

Experience in driver behavior, mobility, or IoT data a plus

Familiarity with AWS SageMaker or cloud deployment preferred

Timeline:
We would like to complete the first version in 2–3 weeks. Follow-up phases may include real-time integration and dashboard deployment.

Budget:
Flexible depending on expertise and scope. Please include:

Estimated delivery time

Budget expectations

Relevant portfolio examples (GitHub, Kaggle, LinkedIn, etc.)

To Apply:
Please submit:

A brief description of your proposed approach

Relevant experience with similar data/modeling tasks

Any initial questions or suggestions for success