Skip to content
View RezaAshrafii's full-sized avatar

Block or report RezaAshrafii

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
RezaAshrafii/README.md

Reza Ashrafi

Statistics at the University of Tehran · Reliable ML and software engineering

I am a Statistics student at the University of Tehran, working on reliable machine learning and applied AI systems. My recent projects study calibration, uncertainty, distribution shift, and sequential decision-making. I also build the software around that work—from experiment pipelines and APIs to interfaces that make results easier to inspect.

I previously studied Electrical Engineering and completed Maktab Sharif's software-development program. I use AI tools throughout development to move faster on implementation, testing, debugging, and documentation, while checking the important decisions against code, data, and reproducible artifacts.

Research work

Project Focus
Calibrated Predictive Reliability A reproducible C-MAPSS study of remaining-useful-life prediction under operating-condition and fault-mode shift, with a focus on calibration, interval reliability, leakage-safe splits, frozen protocols, and artifact-level verification.
PromoGuard Retail Intelligence Evidence-aware promotion auditing on real retail data. It combines time-aware forecast comparison, uncertainty guardrails, explicit abstention, FastAPI contracts, and a Persian Streamlit review interface.
Bearing Prognostics & Value of Information An in-progress research pipeline for sequential bearing-degradation detection, false-alarm control, and inspection decisions. It separates development, calibration, and external NASA IMS validation rather than re-tuning on the final dataset.

Software and AI systems

Project What I built
AURALIS Persian and multilingual meeting intelligence with speech capture, evidence-grounded insights, workspaces, and action tracking. Latest release: v0.10.5 — Audio Path Hardening.
Professor-Aware Exam Coach A local-first study workspace built around course-source retrieval, structured feedback, FastAPI, Next.js, and SQLite.
Tabrizi Bakery A responsive bakery website built with Next.js and TypeScript, with an editorial visual system and a live deployment.

How I work

I am most interested in work that needs both statistical reasoning and solid engineering. I try to define the claim before running an experiment, keep data boundaries explicit, and document what the available evidence does and does not support.

My current stack includes Python, scikit-learn, FastAPI, TypeScript, React, Next.js, SQLite, Streamlit, GitHub Actions, and LLM/RAG tooling. I choose the stack around the problem rather than treating it as the point of the project.

Current direction

My current direction is reliable decision-making for sequential and structured data: calibrated prediction, evaluation under shift, and policies that account for uncertainty before recommending an intervention.

Contact

For research, software, or data/AI product collaboration, reach out through GitHub or LinkedIn.

Pinned Loading

  1. calibrated-reliability calibrated-reliability Public

    Python

  2. promoguard-retail-intelligence promoguard-retail-intelligence Public

    Evidence-aware retail promotion auditing with real-data validation, time-aware forecasting, uncertainty guardrails, FastAPI, and Streamlit.

    Python

  3. bearing-prognostics-voi bearing-prognostics-voi Public

    Sequential bearing degradation detection with calibrated false-alarm control and Value-of-Information inspection policy.

    Python

  4. AURALIS AURALIS Public

    Persian and multilingual meeting intelligence platform with speech capture, grounded insights, workspaces and action tracking.

    JavaScript

  5. professor-aware-exam-coach professor-aware-exam-coach Public

    Local-first AI exam coach for source-grounded academic practice and answer feedback.

    Python

  6. tabrizi-bakery-frontend tabrizi-bakery-frontend Public

    Editorial, responsive heritage bakery website concept built with Next.js, TypeScript and Tailwind CSS.

    TypeScript