I'm an MSc researcher in Computer Engineering (AI) at the University of Genoa,
working as a Research Assistant at CNR within the EU Horizon Europe project REXASI-PRO.

My thesis investigates the reliability of pretrained neural motion prediction models
for autonomous wheelchair navigation — understanding when and why they fail under varying input conditions.

End-to-end Retrieval-Augmented Generation system with a built-in evaluation and failure analysis framework.

• Failure-mode classifier tagging every output as one of 6 interpretable types
  (no_retrieval, wrong_retrieval, hallucination, refusal_when_answerable, partial_answer, ok)
• Retrieval metrics: Precision@k, Recall@k, MRR
• LLM-as-judge faithfulness scoring
• Pluggable backends: OpenAI · Anthropic · local Ollama

Benchmark: Recall@k 0.90 · MRR 0.83 · Latency 2.4ms · 37 tests · CI on every commit

Stack: Python · FastAPI · ChromaDB · Docker · GitHub Actions

Preprint (PDF available) — ETH Pedestrian Dataset

A model-agnostic framework for analyzing failure modes in trajectory prediction systems, evaluated on real-world pedestrian data.

• Input-space sensitivity analysis (orientation–velocity risk regions)
• Interpretable decision tree models for failure rule extraction
• Cross-scene generalization analysis (ETH vs Hotel)

Key Insight: Initial orientation is a dominant global risk factor, while positional features are scene-dependent — indicating limited transferability of failure rules.

📄 Read Paper
📊 Code & Experiments

MSc Thesis — University of Genoa / CNR / REXASI-PRO

Trajectory-level evaluation of pretrained DNN-LNA neural models for autonomous wheelchair navigation.

• Input sensitivity analysis — orientation φ is the dominant risk factor (failures near ±π)
• Goal difficulty mapping — spatial failure patterns across the workspace
• Comparative evaluation of 5 neural architectures (success rates: 25.3% → 99.3%)

Demo:

Trajectory Behaviour Analysis Toolkit

Modular Python toolkit for trajectory risk scoring, failure-case analysis, and interpretable ML explanations.
Includes an interactive Streamlit dashboard — try it live!

• REST API endpoint for trajectory risk scoring (FastAPI)

Safety-Constrained RL for Assistive Robot Navigation

PPO agent with CBF safety layer and curriculum learning for obstacle avoidance navigation.

End-to-end BI workflow — Python, SQL, and Power BI dashboard.

AI & ML — Python · PyTorch · scikit-learn · NumPy · pandas
LLM & RAG — OpenAI API · Anthropic API · ChromaDB · sentence-transformers · FastAPI
MLOps — Docker · GitHub Actions CI · pytest · Prometheus · structlog . Git

🔗 LinkedIn
📧 pouyapd68@gmail.com