Hi! I'm Mocanu Ana-Maria Luisa!

Download CV

About Me

My name is Mocanu Ana-Maria Luisa, but everyone addresses me by the nickname "Medi". I live in Bucharest, Romania.
I am currently a PhD student in Computer Science at the University Politehnica of Bucharest. I also hold a Master's degree in Artificial Intelligence and a Bachelor's in Mechanical Engineering from the same university.
Beyond my research, I teach Python to bachelor's students. Professionally, I work as a Software Engineer at Endava, where I focus on service-oriented and microservice architectures using Java, Spring Boot, and various cloud technologies.

About My Experience

Work Experience

  • Data Scientist - Endava
    • June 2026 - Present

    I apply machine learning and NLP to data-driven problems. This role builds directly on my research work and my software-engineering background from previous fintech projects.

  • Software Engineer - Endava
    • December 2021 - June 2026

    I collaborated with dedicated teams on designing and developing fintech projects. This role enhanced my skills in software development, Agile methodologies, and attention to detail while working with service-oriented and microservice architectures.

  • Teacher - University Politehnica Bucharest
    • October 2025 - Present

    I teach Python to Bachelor's students, helping them with exercises, projects, and understanding core programming concepts. I teach beginner-level Python in Google Colaboratory and review their lab exercises, helping them build practical programming skills.

  • Freelance - Web Developer
    • April 2021 - December 2021

    I created web pages (front-end / back-end and sometimes the design part) in my free time, so I improved my creativity, my efficiency in order to deliver a project faster, my debugging skills and I gained experience in the field.

  • Internship - MARCO & ALEX - Refrigeration Technician
    • July 2020 - August 2020

    I worked mostly as a designer in AutoCAD and other specific softwares to calculate and design according to exact data. This experience trained my quick thinking in unpredictable situations.

Educational Background

  • Ph.D. Candidate in Computer Science
    • University Politehnica Bucharest | 2025 - Ongoing

    Focused on Artificial Intelligence, designing AI models, and applying machine learning techniques to solve complex problems.

  • Master's Degree in Artificial Intelligence
    • University Politehnica Bucharest | 2021 - 2025

    Advanced studies in AI.

  • Bachelor's Degree in Mechanical Engineering
    • University Politehnica Bucharest | 2017 - 2021

    Foundation in engineering and technical problem solving.

Achievements

  • AIDays
    • September 2025

    Poster Presentation

  • Oral Presentation - DAWAK 2026
    • August 2026

    Presented at the 28th International Conference on Big Data Analytics and Knowledge Discovery. Read more about ACAT in the Research section above.

  • Oral Presentation & Poster - CLEF 2026
    • September 2026

    Presented at the Conference and Labs of the Evaluation Forum: IROH ranked 1st on JOKER Task 1 English (0.6347 MAP), and the multimodal sexism-detection system ranked 29th of 114 on EXIST Subtask 2.2. Read more in the Research section above.

  • Poster Session - Best Romanian AI Thesis Awards
    • September 2026

    Presented at Romanian AI Days.

Volunteering Projects

  • Volunteer - Act for Tomorrow
    • July 2019 - September 2019

    Actively contributed to environmental sustainability initiatives by coordinating coastal conservation efforts and community-led waste management. Facilitated educational workshops on ecological responsibility and circular economy practices, strengthening my organizational skills and commitment to community leadership.

  • Volunteer - TOGETHER EVERYONE ACHIEVES MORE (TEAM)
    • 2014 - 2016

    Engaged in community-building initiatives focused on social welfare and youth development. Organized and facilitated interactive workshops and charitable events designed to foster positive social dynamics. This role refined my interpersonal communication, emotional intelligence, and ability to manage complex logistics under tight schedules.

  • Volunteer - Caravana Voluntarilor
    • 2014 - 2015

    Collaborated on large-scale humanitarian projects to deliver community-focused charitable events. Developed a proactive approach to problem-solving and took decisive initiative in high-pressure situations, fostering a deep sense of social accountability and leadership.

Technical Skills

  • Programming Languages

    • Java, Python, JavaScript, HTML, CSS

  • Frameworks & Architecture

    • Spring Boot, Spring, Angular JS, Django, Flask, Microservices, Service-Oriented Architecture, Apache Kafka

  • Databases & Migrations

    • PostgreSQL, MongoDB, Azure Cosmos DB, Redis, MySQL, Amazon DocumentDB, SQLite, Liquibase, Mongock

  • Cloud & DevOps

    • Docker, AWS Services, Microsoft Azure, Google Cloud Platform, Linux, Terraform, GitHub Actions, AWS CodeBuild

  • Artificial Intelligence & ML

    • Machine Learning, Natural Language Processing (NLP), Large Language Models, Multimodal Learning, LLM Fine-tuning (LoRA/QLoRA), PyTorch, Hugging Face

  • Tools & Methodologies

    • Git, GitHub, Bitbucket, Maven, Jira, Postman, Swagger, Bruno, SonarLint, Cucumber, Agile

Publications

Mocanu, A. M. L., Mocanu, S., Truică, C. O., Apostol, E. S. (2026). IROH: Insightful Ranking Of Humor using Multi-Stage Hybrid Retrieval with Rationale-Distilled LLM Judges. In: CLEF 2026, JOKER Lab, Task 1 English. 🏆 1st place · 0.6347 MAP. [Code]

Mocanu, A. M. L., Mocanu, S., Truică, C. O., Apostol, E. S. (2026). Through the Eyes of the Beholder: Biometric and Demographic Conditioning for Multimodal Sexism Detection. In: CLEF 2026, EXIST Lab, Task 2. Top 29 of 114 (Subtask 2.2, soft evaluation). [Code]

Mocanu, A. M. L., Truică, C. O., Apostol, E. S. (2026). ACAT: A Collaborative Platform for Efficient Aspect-Based Sentiment Dataset Annotation. arXiv:2606.04189. [PDF]

Projects

🏆 1st place · CLEF 2026 JOKER IROH: Humor Retrieval Pipeline

BM25 + Dense Retrieval • Cross-Encoder Reranker • LoRA/QLoRA LLM Judges

Built a three-stage humor retrieval pipeline combining hybrid sparse-dense retrieval, a finetuned cross-encoder reranker, and an ensemble of LoRA-adapted LLM judges trained on distilled rationales, ranking 1st on the official JOKER 2026 Task 1 English leaderboard with a MAP of 0.6347.

Top 29 of 114 · CLEF 2026 EXIST Multimodal Sexism Detection in Memes

CLIP + XLM-RoBERTa • FiLM Conditioning • Cross-Attention

Developed a human-centered multimodal model fusing meme text, a generated visual description, the image, annotator demographics, and physiological signals, using cross-attention to ground the description in the image and FiLM to condition on annotator context.

ACAT: Collaborative Annotation Tool for ABSA

Flask • PostgreSQL • Docker

A web-based annotation platform covering four ABSA tasks of increasing granularity, computing inter-annotator agreement (Cohen's/Fleiss' Kappa) directly at export time and exporting to CSV, JSON, and XML. Published as an arXiv preprint.

Scholar Rank: Python Exam IDE

Flask • PostgreSQL • Docker • Monaco Editor

A web-based exam platform where students solve Python problems in a browser IDE and their code is graded automatically inside sandboxed Docker containers. Includes a professor dashboard, randomized problem pools, and anti-cheat measures, and is actively used for live lab exams.

NLP with Disaster Tweets (Kaggle)

Logistic Regression • BERT • BERTTweet Large

Participated in a Kaggle competition to build a machine learning model predicting which Tweets are about real disasters. Achieved a top accuracy score of 0.84554 using a fine-tuned BertTweet Large model, experimenting with various text preprocessing techniques and model architectures.

Abusive Language Detection in Social Media

PyTorch • TensorFlow • Django • Python • Twitter API

Evaluated state-of-the-art abusive language datasets and deep learning models to improve hate-speech classification. Trained 7 different models across 5 distinct datasets to build an enhanced detection algorithm, successfully packaging the models into a full-stack Django web application integrated with the Twitter API.

Department's Website Content Management System

Django • Python • MySQL • JavaScript • HTML5/CSS3

Developed a full-stack Django web application tailored for a university department. Implemented a secure, admin-friendly CMS that enables non-technical academic staff to seamlessly modify pages and manage content.

Bunny NomNom - Android App

Java • Android Studio • Room Database

Built a mobile application to help prevent common pet health issues caused by incorrect alimentation. The app allows users to create specific, healthy meal plans for their bunnies and track their daily diets using a local Room database.

Asame - Social Media for Music Lovers

Django • SQLite • JavaScript • Spotify Web API (Spotipy)

Created a specialized social media platform integrating the Spotify Web API. Implemented full CRUD operations for user posts and designed the backend architecture to support future AI-driven music recommendations based on user taste profiles.

My certifications

Introduction to Model Context Protocol certificate
Introduction to agent skills certificate
Claude Code in Action certificate
Building with the Claude API certificate