Los Angeles, California, United States
I’m a second‑year PhD student in Computer Science at USC’s Viterbi School of Engineering and a Graduate Research Assistant at the Morality and Language Lab, where I explore the intersections of NLP and real‑world deployments. At MOLA Lab, I built CVAT‑BWV, a web‑based annotation platform for the Everyday Respect project, deployed in a secure, on‑premises environment to process sensitive body‑worn camera footage. Designing robust pipelines under strict security constraints has been my daily grind. Simultaneously, I devised a multi‑agent Balderdash simulation to probe creativity and deception in large language models. Our ACL 2024 Workshop paper introduces metrics to evaluate when an LLM is truly ‘inventive’ vs. plausibly bluffing—and whether lying convincingly is smarter than telling the truth. Before my PhD, at USC‑ISI, I directed a midterm‑election Twitter pipeline on AWS/GCP—converting V2 JSON streams into analytics‑ready CSVs—and fine‑tuned BERT/GPT‑2 classifiers for moderation (F1 up to 0.85). I also created an OCR‑based Farsi text annotation tool for historical newspapers and processed a 700 million‑tweet corpus for moral‑sentiment clustering. I’m fluent in Python, and frameworks like PyTorch, Hugging Face transformers, scikit‑learn, plus full‑stack skills in Node.js, React, and cloud/DevOps (AWS, GCP, Docker).
• Build and evaluate supervised classification models on molecular/formulation-structure data, with a focus on graph neural networks (GCNNs) over SMILES-derived molecular graphs. • Create preprocessing pipelines for graph learning (e.g., canonicalization/normalization of SMILES, graph construction + featurization, dataset cleaning, and split strategy). • Prototype internal AI agent workflows for grounded research: orchestrate multi-step retrieval from external APIs/sources and use LLMs to extract normalized structured fields and generate citation-backed summaries. • Implement end-to-end backend services in Python (e.g., FastAPI/Pydantic, agent orchestration such as LangGraph-style DAGs).
I am currently an NLP Research Assistant in the Morality and Language lab.
CSCI 572: Information Retrieval & Web Search Engines (Fall 2025) • Designed grading rubrics for a homework assignment • Graded homework and provided student feedback • Proctored and graded midterm and final exams CSCI 356: Computer Systems (Spring 2026) • Lead weekly discussion sections focused on exam-style problem solving, assignments, and conceptual reinforcement • Hold office hours and mentor students through debugging and systems-level reasoning while promoting independent problem-solving • Contribute to assignment support and course operations, coordinating with the instructional team to ensure consistent guidance and timely communication
Natural language processing research with MINDS group led by Prof. Emilio Ferrara. Using multilingual AI moderators to encourage positive behavior in online social networks.
Worked on my B.Sc. thesis project, which was Information extraction from users’ requests in “SOHA.”
Course Support: - Fundamentals of Computing & Programming (Sep 2017 – Jun 2018) - Advanced Computer Programming (Sep 2017 – Jun 2018; Head TA Aug 2018 – Jan 2020) - Data Structures (Sep 2019 – Feb 2020) - Operating Systems (Sep 2019 – Feb 2020; Feb 2021 – Jun 2021) - Internet Engineering (Feb 2020 – Jul 2020; Sep 2020 – Jun 2021) - Artificial Intelligence & Expert Systems (Feb 2021 – Jun 2021) Leadership & Mentorship: - As Head TA for Advanced Computer Programming, led a team of 5+ assistants: coordinated lab schedules, standardized grading rubrics, and liaised directly with faculty. - Mentored undergraduate students one-on-one during office hours and lab sessions, improving average course evaluation scores by 15%. Instructional Duties: - Designed and delivered weekly recitation materials, coding demonstrations, and in-class quizzes. - Managed assignment and exam grading (1000+ submissions per term), provided detailed feedback, and held review workshops. - Assisted professors with syllabus design, lecture preparation, and integration of new pedagogical tools. Faculty Collaborations: Worked under the guidance of Dr. Mojtaba Vahidi Asl, Dr. Mehrnoush Shamsfard, Dr. Alireza Shameli, Dr. Parham Alvani, Dr. Seyed Ehsan Mahmoudi, Dr. Mohsen Ebrahimi Moghadam, Dr. Monire Abdus, Dr. Sadegh Aliakbari & Dr. Dorri.