Project overview

• In project management, one major challenge is realistically estimating the time, cost, and resources required for each task. Traditional methods (PERT, Gantt, manual planning) often rely on the project manager’s experience and can be biased or inaccurate.
• The project aims to develop a web application that automatically estimates project tasks based on historical project data and AI predictions.
• Objectives: develop an intelligent web application for task estimation and use AI/ML to predict duration, cost, and workload for each task.
• Reference: PRX-2026-12 — Intelligent Project Task Estimation Application.

Technical scope & technologies

• Frontend: React JS for building the user interface and interactive estimation dashboards.
• Backend: Node JS to serve APIs, integrate ML inference, and handle business logic.
• Data storage: PostgreSQL for storing historical project data, task metadata, predictions and user information.
• Machine learning: design and train models to predict task duration, cost and workload from historical project datasets; evaluate and validate model accuracy and uncertainty.

Trainee responsibilities

• Collect and preprocess historical project data (cleaning, feature engineering, labeling) and define relevant metrics (duration, cost, workload).
• Develop ML pipelines: prototype, train and validate predictive models (regression, ensemble methods, or deep learning as appropriate) to estimate task parameters.
• Implement backend integration: create RESTful APIs in Node.js to serve model predictions and handle data interactions with PostgreSQL.
• Implement frontend features in React: estimation forms, visualization of predicted durations/costs/workloads, comparison with historical data and manual overrides.

Expected deliverables

• A working web application (React frontend + Node.js backend) that produces task-level estimations from uploaded or stored historical data.
• Trained ML models and reproducible training pipeline, plus evaluation reports (metrics, validation results, error analysis).
• Database schema and populated sample dataset in PostgreSQL, plus scripts for data ingestion and preprocessing.
• Documentation: installation/deployment steps, user guide, technical report describing methodology and model choices.

Profile & required skills

• Engineer profile / Bachelor’s degree (required) — 1 Trainee position.
• Proficiency with React JS, Node JS and relational databases (PostgreSQL); experience in building full-stack web applications.
• Knowledge of machine learning concepts and practical experience with ML libraries (scikit-learn, TensorFlow or PyTorch), data preprocessing and model evaluation.
• Good software engineering practices: version control (Git), testing, API design and basic deployment knowledge.

How to apply

• Apply via the trainees platform: https://trainees-platform.proxym-group.net
• Include in your application details on past projects or coursework in web development and machine learning, and any sample code or portfolio relevant to project estimation or data-driven applications.