Project overview

• An applied AI project focused on designing an intelligent assistant that supports complex, human-centric decision processes using structured information and advanced reasoning.
• Goal is to build a robust, end-to-end prototype (frontend + backend + AI layer) that demonstrates practical value in a realistic HR workflow use case.

Missions / Objectives

• Design an intelligent assistant that helps professionals make better, faster decisions using structured information and AI-driven insights.
• Explore how advanced language models can be guided and constrained to behave reliably in complex, real-world workflows and produce explainable outputs highlighting key points and evidence.
• Model domain knowledge so AI can use it consistently while remaining understandable to non-technical users.
• Evaluate the assistant’s behaviour against simpler baselines to measure improvements in structure, relevance and usability.

Technical scope & deliverables

• Build a full-stack prototype: frontend (React.js), backend (Node.js), database (MySQL + vector DB), caching (Redis), and containerization (Docker).
• Integrate Gen-AI components: SentenceTransformers, LangChain patterns, vector search, and produce clear, explainable model outputs instead of opaque answers.
• Deliverables include a working prototype, documentation of design and evaluation, comparison metrics versus baselines, and user-facing explanation interfaces.

Expected skills & activities

• Strong software development skills in React.js and Node.js, experience with databases (MySQL) and vector databases, and familiarity with Docker and Git.
• Experience or willingness to work with Gen-AI tooling (SentenceTransformers, LangChain), Redis for caching, and testing frameworks (Jest) is required.
• Tasks include data & domain modelling, prompt/chain engineering, backend/frontend integration, evaluation design, and user-facing explanation design.

Logistics

• Internship duration: 4-6 months.
• The project emphasizes reproducible evaluation, explainability, and practical integration into HR decision workflows.

Evaluation & success criteria

• Success measured by improvements in structure, relevance and usability compared to simpler baselines.
• Qualitative feedback from realistic user scenarios and quantitative metrics from evaluation experiments.