I'm a PhD candidate in Computer Science at the University of California, Davis. My work revolves around applying machine learning techniques, especially from the natural language processing community, to software engineering problems.
🧠 Expertise: Machine Learning, Research, Python, ML4SE, AI4Code
🔗 [Personal website] - [LinkedIn]
🎓 Education
- PhD Candidate in Computer Science, Advised by Prem Devanbu at University of California, Davis (Dec. 2026)
- Bachelor in Computer Science & Engineering at Free University of Bozen-Bolzano (Sep. 2019)
- Computer Science Study Abroad at College of Charleston (Dec. 2017)
👨💻 Experience
- Research Scientist Intern at IBM Research
- Developed AutoPDL, a source-to-source optimizer for Prompt Declaration Language (PDL) that automatically searches and optimizes prompt programs over user-defined spaces.
- Implemented a library of agentic reasoning patterns (ReAct, ReWOO) and conducted empirical evaluation across 7 LLMs and multiple benchmarks, achieving +9 pp average accuracy over baselines. Published AutoML'25 (1st author).
- Graduate Student Researcher at UC Davis
- Investigated confidence vs correctness (calibration) of LLMs for code eg GPT3.5 in program synthesis, line completion, and defect detection tasks. Built dataset of Python functions for prompting & evaluating model completions. Published ICSE'25 (1st author).
- Designed and trained custom HuggingFace/PyTorch BERT model for embedding dynamic analysis artifacts for source code retrieval using FAISS and evaluated performance. Built dataset of open source projects under instrumentation. MS Thesis. Published FSE'23 SRC (1st author).
- Teaching: ECS 160 Software Engineering, ECS 171 Machine Learning: Held discussion section & office hours, produced & graded homework, proctored & graded exam, answered student questions on discussion forum
- Machine Learning Intern at Overstock
- Built & deployed app to analyze product embedding spaces using nearest neighbor visualizations, linear & classification probes, and catalog coverage analysis for recommender systems ML team, optimizing their development and debugging workflow.
- Software Engineer at Floryn BV
- Developed systems to monitor and interpret ML model outputs, giving critical decision insights to technical & non-technical stakeholders e.g. through the use of SHAP.
🛠️ Skills
- Programming Languages: Python, PyTorch, Ruby, C, SQL (Postgres), Java, TypeScript
- Developer Tools: Git, Docker, Terraform, AWS, CircleCI, VS Code
- Libraries: pandas, NumPy, Huggingface, Matplotlib, plotly