ML/AI Engineer | PhD Mathematician

Former Postdoc Researcher at KTH Royal Institute of Technology | Former PiSchool AI Fellow | Based in Rome

I'm transitioning into ML engineering with hands-on experience building real-time AI systems. My background in pure mathematics (PhD in Algebraic Geometry) gives me a strong foundation for solving complex ML problems. I've built a low-latency speech translation pipeline and contribute to formal verification projects in Lean4.

Actively seeking ML/AI Engineering roles in Rome or remote. Available to start before February 2026.

ML & AI Frameworks PyTorch • TensorFlow • Hugging Face • XGBoost • SHAP

Programming & Infrastructure Python • Git • Linux • Docker • Google Cloud Platform

Data & APIs MySQL • PostgreSQL • FastAPI • Streamlit

Specialized Skills

• Real-time speech translation and NLP
• Financial ML (sentiment analysis, fraud detection)
• Real-time audio processing (sounddevice, VAD, streaming)
• Mathematical modeling and optimization
• Formal verification (Lean4)

Production-ready Italian-English translation pipeline for technical lecture environments

• BLEU ≥40, COMET ≥0.75 on scientific content with ≤4s latency
• Complete streaming pipeline: Whisper ASR → Translation → Kokoro TTS
• Team leadership role in agile development environment
• Technologies: Whisper, FastWhisper, Kokoro TTS, Qwen2.5, PyTorch, CUDA

Production-ready sentiment analysis for financial text using fine-tuned FinBERT

• 79% accuracy with optimized negative sentiment detection (61% improvement)
• FastAPI backend with Docker deployment and comprehensive testing
• Caching, batch processing, and monitoring capabilities

End-to-end fraud detection pipeline with model interpretation

• XGBoost classifier with 99% AUC and 80% PR-AUC
• SHAP integration for model explainability
• Interactive Streamlit demo for real-time predictions

Formalizing mathematical foundations in Lean4 proof assistant

• 700+ lines of verified code for topological Krull dimension theory
• Part of the code was successfully integrated into mathlib (Lean's core math library): Krull-PR
• Technologies: Lean4, formal verification, type theory

PhD in Mathematics (Algebraic Geometry) | Postdoc Researcher at KTH | Former PiSchool AI Fellow

Currently based in Rome and actively looking for ML/AI engineering opportunities (remote or Rome-based). Available to start before February 2026.

LinkedIn: alessandro-d-angelo

Academic Website: sites.google.com/view/alessandro-dangelo