cv

M.Sc.

University of Tehran

IT Engineering

• Thesis: Design and Implementation of an Intelligent Assistant for Clinical Assessment and Diagnosis of Mental Disorders Using Large Language Models

B.Sc.

University of Isfahan

Computer Engineering, IT Major

• Final Project: Design and Development of a Structured Feedback Collection and Evaluation System for University Teaching and Research Assistants to Enhance Instructional and Research Support Quality

SchemaGraphSQL: Efficient Schema Linking with Pathfinding Graph Algorithms for Text-to-SQL on Large-Scale Databases

Accepted at EACL 2026

Safdarian, A., Mohammadi, M., Jahanbakhsh, E., Naderi, M. S., & Faili, H.

UT-NLP at SemEval-2025 Task 11: Evaluating Zero-Shot Capability of GPT-4o mini on Emotion Recognition via Role-Play and Contrastive Judging

SemEval 2025

Mohammadi, M.*, Safdarian, A.*, & Faili, H. (*Equal contribution)

PerMedCQA: Benchmarking Large Language Models on Medical Consumer Question Answering in Persian Language

Jamali, N., Mohammadi, M., Baledi, D., Rezvani, Z., & Faili, H.

Ranked 1st nationwide in IT Engineering, Graduate Entrance Exam (Konkur)

Iran National Graduate Admissions

Ranked first nationwide in IT Engineering in Iran's national graduate entrance examination.

Senior AI (Research) Engineer (Part-time, Remote)

IntRest

End-to-end ownership of an AI microservice for food analysis, covering agentic system design, retrieval, backend architecture, and production integration.

• Own the end-to-end AI microservice for food analysis: refactored the full codebase, built an agentic RAG system over food and ingredient data for accurate and efficient analysis, and lead the software layer from requirements and backend design to API endpoints and production integration.

AI Software Engineer (Part-time)

Adan AI

Production-oriented LLM applications, agentic customer-service systems, Persian document RAG pipelines, FAQ-bot automation, and reusable LangGraph infrastructure.

• Led architecture and delivery of an agentic customer-service chatbot for a national bank: translated PM requirements into user stories; designed a ReAct/Reflexion workflow with more than 15 tool steps; implemented and reviewed LangGraph and OpenAI modules; defined and tracked evaluation metrics for production.
• Built an end-to-end PDF-to-chunking pipeline for Persian documents, including table/chart-aware parsing and entity extraction; reduced processing cost by 50% and improved retrieval accuracy by 8% on a human-annotated RAG evaluation set.
• Created an automatic FAQ-bot generator that ingests Excel FAQs, cleans and synthesizes data from seed examples, and deploys a ready-to-use chatbot, cutting onboarding time from days to less than one hour with no manual coding.
• Engineered reusable agent-infrastructure modules for session-level user-token management, conversation summarization, entity enrichment, and metadata handling; these modules were adopted across LangGraph projects, eliminating duplicate code and accelerating new-bot delivery.

Chief Teaching Assistant

Natural Language Processing

Chief TA for the Spring 2026 NLP course at the University of Tehran.

• Managed the TA team in designing and grading course assignments, organized class quizzes, and coordinated and scored student paper presentations.

Chief Teaching Assistant

Natural Language Processing

Chief TA for the Fall 2025 NLP course at the University of Tehran.

• Led a 12-member TA team and coordinated course planning, weekly schedules, quizzes, and exams.

Workshop Instructor

Artificial Intelligence & Digital Transformation

Workshop instructor for a university-wide course on AI and digital transformation.

• Delivered two large-venue lectures to more than 500 students in the course's inaugural university-wide offering open to all fields.
• Covered prompting, language-agent concepts, and building simple workflow agents in n8n, tailored for non-engineering cohorts.

Graduate Teaching Assistant

Large Language Models

Graduate TA for a course on large language models, language agents, and applied LLM systems.

• Designed a multi-stage course assignment on language agents, from prompt-router baselines to ReAct-style agents solving Text-to-SQL on the BIRD dataset.
• Authored the assignment rubric and graded the assignment, midterm, and final.
• Led a hands-on workshop on building agents with LangGraph.

Graduate Teaching Assistant

Deep Learning & Neural Networks

Graduate TA for deep learning and neural networks, with emphasis on LLM architecture and fine-tuning.

• Designed the course project on LLM fine-tuning: Gemma 2 for Persian chat via Prefix-Tuning, LoRA, and QLoRA.
• Delivered a workshop on LLM architecture and fine-tuning with Hugging Face Transformers.

Graduate Teaching Assistant

Intelligent Information Retrieval

Graduate TA for information retrieval, sparse retrieval, dense retrieval, and RAG pipelines.

• Designed NanoSparseRetriever, a from-scratch Python assignment covering indexing, query/document processing, and pluggable scoring.
• Students implemented TF-IDF and BM25 and probed parameter effects.
• Taught a project workshop on OpenAI API and RAG pipelines using LanceDB and open-source dense retrievers.

Graduate Teaching Assistant

Natural Language Processing

Graduate TA for NLP, focused on agentic NLP systems, RAG, and chatbot project design.

• Authored the final project ChatFood: an agentic food-ordering chatbot with RAG over a food book, tool use for order/track/cancel operations, and a Reflexion-based recommender.
• Provided runnable backend and database starter code for the final project.
• Ran a 4-hour workshop on chatbots and agents with LangGraph, LanceDB, and Chainlit, covering design patterns and evaluation.

Teaching Assistant

Undergraduate Computer Engineering Courses

Teaching assistant for multiple undergraduate computer engineering courses.

• Courses: Algorithm Design, Discrete Mathematics, Programming Foundations, IT Foundation, and E-Commerce.

Text-to-Insight Data Analyst Agent (Text2SQL & Data Visualization) · GitHub

End-to-end agent converting natural-language questions into SQL over a curated SQLite database from the BIRD dataset, then auto-generating Python analysis code for tables/plots and a written report.

• Built a natural-language-to-SQL workflow over a curated SQLite database.
• Generated Python analysis code for tabular analysis and visualization.
• Produced written analytical reports from query results.
• Included an LLM-as-judge framework to evaluate component quality and end-to-end answers, including report accuracy and SQL exact match.

MedYar: Health Reservation Agent (Triage & Booking) · GitHub

Doctor-visit assistant performing symptom intake and specialty routing, collecting user preferences, checking availability, and booking appointments on Paziresh24.

• Implemented symptom intake and specialty routing.
• Collected user preferences such as physician and time.
• Checked appointment availability and completed booking via web automation.
• Implemented as a ReAct-style tool-use agent with a full booking flow.
• Reported successful end-to-end demo execution.

Committee Member

ACM Student Chapter, University of Isfahan

Organized programming contests, taught algorithm design, and participated in weekly ICPC trainings.

• Organized programming contests and educational sessions.
• Taught algorithm design to student participants.
• Participated in weekly ICPC training activities.

Team Lead & Organizer

TEDxUniversityOfIsfahan

Founded TEDxUniversityOfIsfahan in 2019 and led end-to-end event organization.

• Founded TEDxUniversityOfIsfahan in 2019.
• Led planning, budgeting, sponsorship, marketing, and speaker coaching.
• Helped establish a student-led event series that now runs annually through the student community.
• Supported talks that reached thousands of views.