Data Analyst · AI Engineer · M.Sc. AI Student

I turn messy data into clear decisions—and AI ideas into systems that actually work.

I work across the full data journey: extraction, validation, analysis, modeling, visualization, and communication—then connect the right ideas to APIs and automation when the problem calls for it.

Explore my work Download resume
GitHub LinkedIn Email
Portrait of Mobina Babarashani
RAG
ML
API
4Decision-focused portfolio projects
18.90M.Sc. GPA out of 20
1st & 2ndNational automation awards

About

From raw data to a decision someone can use.

I am an Artificial Intelligence master's student with a background in Mechanical Engineering and hands-on experience in data analysis. I care about the full path from raw records to a trustworthy conclusion: cleaning, validation, exploratory analysis, KPI design, modeling, visualization, and clear communication.

My portfolio spans e-commerce operations analytics, a RAG-powered operations agent, customer-retention intelligence, and wind-turbine performance modeling. I build with Python, SQL, Pandas, FastAPI, n8n, Docker, and machine-learning tooling—and keep the business question visible throughout the technical work.

Data Analysis & KPI Design Applied Machine Learning RAG & LLM Workflows API & Workflow Automation

Featured projects

Systems built around real decisions.

All repositories ↗
fulfillment_analysis.py

01 validate(order_lifecycle)

02 calculate(on_time_delivery)

03 segment(delay_patterns)

04 translate(findings → actions)

OUTPUTOperational insight, not just a chart.
01

Operations Analytics · KPI Design · Confidential Data

E-commerce Fulfillment Analytics

An end-to-end operational analysis of a real e-commerce order lifecycle, built to reveal delivery performance, process bottlenecks, and actionable opportunities for improvement.

  • Cleaned, validated, aggregated, and explored shipment-level data with Python, Pandas, and SQL.
  • Designed operational KPIs including On-Time Delivery using business-defined delivery rules.
  • Investigated timestamps, shipment states, and process patterns to explain delays—not merely count them.
PythonPandasSQLEDAKPI Analysis

Dataset and implementation details remain private due to confidentiality.

n8n workflow connecting Telegram, FastAPI, Google Sheets, and escalation logic
02

RAG · Automation · Backend

SupportOps AI Agent

A knowledge-grounded operations workflow that receives Telegram messages, retrieves controlled context from ChromaDB, generates an answer through OpenRouter, and routes low-confidence or policy-sensitive cases to a tracked escalation path.

  • Orchestrated Telegram, n8n, FastAPI, ChromaDB, OpenRouter, and Google Sheets.
  • Added confidence scoring, category and urgency classification, ticket IDs, and escalation rules.
  • Included API tests, workflow templates, screenshots, setup documentation, and explicit production limitations.
PythonFastAPIn8nChromaDBDocker
View repository ↗
Campaign targeting efficiency chart for customer churn risk ranking
03

Machine Learning · Business Intelligence

Anticipate & Act

An end-to-end churn decision system that ranks customers by risk and converts model probabilities into practical retention actions instead of stopping at binary predictions.

0.8407ROC-AUC
XGBoostFinal model
scikit-learnXGBoostStreamlitPandas
View repository ↗
Wind turbine power calculator diagram showing rotor diameter, swept area, wind speed, and power output
04

Industrial AI · Regression Analytics

Wind Turbine Performance Analytics

A SCADA-based machine-learning study that estimates expected turbine output and translates performance gaps into maintenance-oriented energy KPIs.

0.9858Best R²
75.41 kWBest MAE
Gradient BoostingSCADANumPyMatplotlib
View repository ↗

Technical toolkit

What I use to build.

01

Analysis & Insight

Data cleaning, exploratory analysis, KPI engineering, statistical thinking, dashboarding, and translating structured data into clear business recommendations.

SQLPandasNumPyExcelPower BI
02

Machine Learning & AI

Feature engineering, regression and classification, model evaluation, RAG pipelines, embeddings, vector retrieval, and LLM API integration.

scikit-learnChromaDBMLflowRAG
03

Backend & Automation

REST API development, webhooks, workflow orchestration, bot integration, structured validation, decision logging, and containerized services.

PythonFastAPIn8nDockerPostman
04

Systems & Delivery

Linux and networking fundamentals, Git-based collaboration, relational database basics, scheduling concepts, and practical technical documentation.

LinuxGit/GitHubPostgreSQLAirflowdbt

Experience

Professional background.

Aug 2024 — Present

AI Engineer / Data Analyst

Freelance

Deliver AI and data-analysis projects using Python, SQL, and automation tools, translating business requirements into focused, practical, data-driven solutions.

Jul 2023 — Aug 2024

Technical Specialist (R&D)

Pichka Industrial Group

Worked with CRM and structured internal data, supported operational reporting and follow-up workflows, documented technical handoffs, and identified Python/API automation opportunities.

Education

Academic foundation.

2025 — Present

M.Sc. in Artificial Intelligence

University of Science and Culture, Tehran

GPA: 18.90 / 20
2019 — 2023

B.Sc. in Mechanical Engineering

Imam Khomeini International University

GPA: 16.02 / 20

Recognition

National Industrial Automation Awards

1st place in Phase 2 and 2nd place in Phase 1 at the First National Shahid Babaei Event, supported by Iran's National Elites Foundation.

Continuous learning

Strong foundations. Always expanding.

Professional working English · Recommendation available from Pichka Industrial Group

01

Linux & NetworksLPIC-1 (101–102), Linux Essentials, CompTIA Network+

02

Data & AIPandas, AI Ethics, Deep Learning, FastAPI for LLM Applications

03

Engineering PracticeGit & GitHub fundamentals, Web Development Bootcamp

Contact

Let’s build something useful.

I am interested in data-analysis and applied-AI opportunities where rigorous thinking, useful automation, and clear communication all matter.

Send an email Download resume