Driver Behavior Scoring ML Model Development
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
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