AI Specialist
Payne Lettings
AI Specialist
£65000
Payne Lettings, Abertawe
- Full time
- Permanent
- Remote working
Posted 6 days ago, 23 Jun | Get your application in now to be included in the first week's applications.
Closing date: Closing date not specified
job Ref: 8bcc3195a4414c6f833e35996e423533
Full Job Description
- Collaborate with stakeholders to define the problem and establish clear objectives for predicting and locating risks in construction cashflows.
- Determine the scope and desired output, such as risk scores or classifications for different construction scenarios.
- Identify and access data sources, including historical project data, weather data, economic indicators, geographic information, and project management data.
- Gather data from APIs, public datasets, and web scraping.
- Clean and preprocess data to handle missing values, standardize data, and ensure consistency.
- Perform feature engineering to create new features from raw data and enhance model input.
- Integrate different datasets into a single, coherent dataset, aligning time-series data correctly.
- Select initial models, focusing on ensemble methods like Random Forest and Gradient Boosting Machines.
- Prepare temporal data for time series analysis, considering models like ARIMA for initial analysis.
- Split data into training and testing sets, using cross-validation to ensure robust evaluation.
- Train initial models and evaluate their performance using validation metrics.
- Train ARIMA models and consider LSTM networks for complex temporal patterns.
- Compare model performance using validation metrics and cross-validation results.
- Present model predictions to domain experts for validation and refine predictions based on feedback.
- Implement ensemble methods and stacking to improve overall performance.
- Analyse feature importance to identify key predictors and refine features to enhance model performance.
- Continuously iterate on feature engineering based on feature importance analysis.
- Integrate the trained model into the web application, using RESTful APIs for real-time predictions.
- Monitor model performance using live data and implement a feedback loop for retraining and refinement.
- Incorporate human oversight, allowing users to provide feedback on model predictions.
- Regularly retrain the model with new data and use automated retraining processes within the CI/CD pipeline.
Bachelor's or Master's degree in Computer Science, Data Science, AI, or a related field. - Proven experience in AI/ML model development and deployment, particularly with ensemble methods and time series analysis.
- Proficiency in JavaScript, Node.js, Express.js, and React for developing web applications.
- Strong understanding of data collection, pre-processing, and feature engineering techniques.
- Ability to work collaboratively with stakeholders and domain experts to refine model predictions.
- Excellent problem-solving skills and attention to detail.