Artificial Intelligence Intern
2 giorni fa
Paullo, Lombardy, Italia
Cambrex
Tempo pieno
Gratuito con email o Google
Salva questo lavoro e mantieni la tua ricerca organizzata
Crea un account gratuito per salvare lavori, creare avvisi e tornare a questa inserzione dalla tua dashboard.
Gratuito con email o Google
Continuando accetti i nostri Termini & Informativa sulla privacy.
Cambrex Profarmaco Milano is looking for a curious and dynamic intern interesed in Artificial Intelligence.
For our Site in Paullo (Milan) we are looking for a trainee interested in with an expertise in ML methods and AI to insert in our Analytical Development team for an industrial project focused on building an AI-powered HPLC recommendation engine. This project leverages proprietary pharmaceutical data to develop machine learning models that predict optimal chromatographic conditions for new compounds.
Responsibilities:
• Design and curate a structured experimental knowledge database linking molecular representations (SMILES, InChI, fingerprints), physicochemical descriptors, chromatographic method parameters, and experimental outcomes, with a focus on data quality, reproducibility, and design suitable for ML;
• Develop a feature selection pipeline for classical descriptors (RDKit, Mordred) and learned representations (molecular fingerprints, graph-based embeddings);
• Research, train, and benchmark predictive models for chromatographic outcomes (retention behaviour, mobile phase strength, column selectivity class), exploring both interpretable models and state-of-the-art approaches;
• Design a molecular similarity module grounded in chemical space geometry, evaluating distance metrics and embedding spaces for nearest-neighbour method retrieval from historical data;
• Build a recommendation engine that unifies predictive modelling and the similarity module, with uncertainty quantification, confidence scoring, and explainability to support trust and adoption by domain scientists;
• Extend the engine into an agentic LLM interface that allows natural language interaction with the underlying models and database;
• Validate the system on held-out experimental data, document methodology to publication standard, and present research outputs to both technical and domain-expert audiences. Qualifications and
Skills:
• Degree in Computer Science/Engineering or a closely related field;
• Deep understanding of ML and AI.
• Strong research instincts: ability to identify the right problem formulation, design-controlled experiments and critically evaluate model behaviour rather than just benchmark metrics;
• Proficiency in Python and relevant libraries (scikit-learn, pandas, NumPy, PyTorch, TensorFlow), comfort reading and adapting research code. Soft
Skills:
• Good interpersonal and communication skills.
• Ability to work effectively in a team and flexibility;
• Proactivity, a strong focus on results, and problem-solving skills;
• Ability to work independently and communicate technical results to a non-specialist audience. It would be considered a plus:
• Familiarity with molecular representations and cheminformatics tools (SMILES, fingerprints, graph neural networks for molecules) or willingness to learn;
• Active interest in explainable and interpretable ML (XAI), particularly in applied scientific contexts where trust and transparency are critical;
• Hands-on experience with LLM tool-use, function calling, or agentic frameworks or conceptual grounding in how LLMs interact with external systems;
• Exposure to scientific, industrial, or experimental datasets with inherent noise, class imbalance, or sparse labelling, common in real-world R&D settings. What You Will Gain:
• Access to a proprietary industrial HPLC dataset not available in academic settings;
• AI research challenge with real-world constraints;
• Immersion in cheminformatics and pharmaceutical analytical R&D, with direct collaboration with domain scientists who will challenge and sharpen your modelling decisions;
• Research outputs aligned with your doctoral trajectory: publishable methodology, a working system demonstrating scientific AI in industrial settings;
• Mentoring from both chemoinformatics and domain experts, with genuine intellectual exchange.
Location:
Cambrex Profarmaco Milano Srl, Paullo (MI) On-site or Hybrid Model (to be discussed) Contract:
We offer
1 year intership, details will be clarified during the interview process.
Responsibilities:
• Design and curate a structured experimental knowledge database linking molecular representations (SMILES, InChI, fingerprints), physicochemical descriptors, chromatographic method parameters, and experimental outcomes, with a focus on data quality, reproducibility, and design suitable for ML;
• Develop a feature selection pipeline for classical descriptors (RDKit, Mordred) and learned representations (molecular fingerprints, graph-based embeddings);
• Research, train, and benchmark predictive models for chromatographic outcomes (retention behaviour, mobile phase strength, column selectivity class), exploring both interpretable models and state-of-the-art approaches;
• Design a molecular similarity module grounded in chemical space geometry, evaluating distance metrics and embedding spaces for nearest-neighbour method retrieval from historical data;
• Build a recommendation engine that unifies predictive modelling and the similarity module, with uncertainty quantification, confidence scoring, and explainability to support trust and adoption by domain scientists;
• Extend the engine into an agentic LLM interface that allows natural language interaction with the underlying models and database;
• Validate the system on held-out experimental data, document methodology to publication standard, and present research outputs to both technical and domain-expert audiences. Qualifications and
Skills:
• Degree in Computer Science/Engineering or a closely related field;
• Deep understanding of ML and AI.
• Strong research instincts: ability to identify the right problem formulation, design-controlled experiments and critically evaluate model behaviour rather than just benchmark metrics;
• Proficiency in Python and relevant libraries (scikit-learn, pandas, NumPy, PyTorch, TensorFlow), comfort reading and adapting research code. Soft
Skills:
• Good interpersonal and communication skills.
• Ability to work effectively in a team and flexibility;
• Proactivity, a strong focus on results, and problem-solving skills;
• Ability to work independently and communicate technical results to a non-specialist audience. It would be considered a plus:
• Familiarity with molecular representations and cheminformatics tools (SMILES, fingerprints, graph neural networks for molecules) or willingness to learn;
• Active interest in explainable and interpretable ML (XAI), particularly in applied scientific contexts where trust and transparency are critical;
• Hands-on experience with LLM tool-use, function calling, or agentic frameworks or conceptual grounding in how LLMs interact with external systems;
• Exposure to scientific, industrial, or experimental datasets with inherent noise, class imbalance, or sparse labelling, common in real-world R&D settings. What You Will Gain:
• Access to a proprietary industrial HPLC dataset not available in academic settings;
• AI research challenge with real-world constraints;
• Immersion in cheminformatics and pharmaceutical analytical R&D, with direct collaboration with domain scientists who will challenge and sharpen your modelling decisions;
• Research outputs aligned with your doctoral trajectory: publishable methodology, a working system demonstrating scientific AI in industrial settings;
• Mentoring from both chemoinformatics and domain experts, with genuine intellectual exchange.
Location:
Cambrex Profarmaco Milano Srl, Paullo (MI) On-site or Hybrid Model (to be discussed) Contract:
We offer
1 year intership, details will be clarified during the interview process.