morteza.
Open to internships & thesis collaborations

Morteza Heidari.

Machine Learning & Data Engineering

Sensor data · time-series forecasting · AI/data pipelines — built for real projects.

I am a Machine Learning and Data Engineering student based in Rovaniemi, Finland. I work on practical data and AI projects — sensor data analysis, forecasting, OCR, NLP pipelines, dashboards, and backend systems. I enjoy turning messy real-world data into clear findings and working applications.

  • Python
  • pandas
  • PyTorch
  • scikit-learn
  • Time-series
  • SQL

Currently exploring:LLM fine-tuning·Time-series forecasting·FastAPI on Docker

See projectsGet in touch Rovaniemi, Finland
Portrait of Morteza Heidari
available
Featured projects
6
Domains
SensorForecastingOCRNLPWeb
Based in
Rovaniemi, FI
01 — about

Data first, then the model.

I like working with data because it shows the real situation behind a problem. In my projects I usually start by asking simple questions: What data do we have? What is missing? What can we trust? What can we not conclude yet? What solution would actually help?

This way of working has shaped my projects on heavy-vehicle sensor data, road-weather analysis, forecasting, OCR pipelines, and AI-assisted reporting. I try to keep my work practical, honest, and useful enough that another person can understand, test, and improve it.

01

Honest data

State what the data shows, what it does not, and where the gaps are — before drawing conclusions.

02

Simple solutions

Prefer the smallest model, the shortest pipeline, and the clearest dashboard that actually solves the problem.

03

Reproducible work

Versioned data, pinned dependencies, and notebooks that re-run end-to-end without hidden state.

02 — experience · 5

Applied work & internships.

  1. Project Worker — AI Accelerator project

    10/2025 — 02/2026

    Lapland University of Applied Sciences · Rovaniemi, Finland

    • Komee E-commerce — sales forecasting and cash-flow modelling for inventory and planning decisions.
    • Sydämen-Nosté Dashboard — FastAPI + React app that scrapes company sites, extracts contacts with AI, stores results in Google Sheets.
    • Video OCR + Audio Report Pipeline — turns pipe-inspection videos into structured CSV reports.

  2. Practical trainer

    06/2025 — 08/2025

    Lapland UAS / AI Accelerator · Rovaniemi, Finland

    • Built a news-monitoring agent that collects, filters, and summarises updates about specific companies (BiSniffer).

  3. Internship — AI.R Arctic AI & Robotics

    10/2024 — 03/2025

    Lapland UAS · Robotics Lab · Rovaniemi, Finland

    • Built an Arctic winter dataset using thermal, RGB, and LIDAR sensors as part of a collaborative project.

  4. Summer Intern — AI.R Arctic AI & Robotics

    05/2024 — 08/2024

    Lapland UAS · Robotics Lab · Rovaniemi, Finland

    • ESP32 / Arduino Nano implementation; switchable RC and autonomous modes.
    • Wrote functions in C and worked with ROS2.

other experience

Social Media Ambassador

09/2025 — 05/2026

University of Lapland Rovaniemi, Finland

Present student life in Rovaniemi on Instagram for the university.

03 — skills

What I work with.

Tools I actually use in projects. Items marked with core are part of my primary day-to-day stack. Each card shows how many of the 11 projects below use a tool from that group.

Data analysis & preparation

used in 9/11 ↓
  • Python
  • pandas
  • Data cleaning
  • EDA
  • Missing values
  • Outliers
  • Feature engineering
  • Sensor data

Machine learning & AI

used in 6/11 ↓
  • scikit-learn
  • TensorFlow
  • PyTorch
  • Time-series forecasting
  • Prophet
  • Monte Carlo
  • NLP
  • OCR
  • Computer vision
  • LLM fine-tuning
  • Hugging Face

Web & backend

used in 5/11 ↓
  • FastAPI
  • Flask
  • React
  • Vite
  • SQL
  • SQLite
  • MySQL
  • Docker
  • REST APIs
  • MQTT

Applied work

used in 2/11 ↓
  • Dashboards
  • Data pipelines
  • Web scraping
  • Playwright
  • Forecasting tools
  • Route-risk backends
  • AI-assisted reporting
04 — featured projects · 6

Practical, honest, useful.

Hazard Situation Analysis & Prediction

Lapland UAS

01/2026 — 04/2026
Sensor

Project manager · data analyst · backend developer

Heavy-vehicle sensor data from a Finnish logistics case: Flic button presses, Ruuvi DF5 motion tags, and MD30 road-weather data. Cleaned and merged the datasets and studied missing data and sensor gaps.

outcome → Identified root causes why hazard presses cannot be classified from current data. Proposed concrete sensor and pipeline fixes. Built a FastAPI route-risk backend combining route, road, weather, and traffic signals.

view case study

what i learned → When the label rate is too low and sensor coverage is uneven, the honest output is a data-quality report and a plan for the next collection round — not a forced classifier.

  • Python
  • pandas
  • FastAPI
  • Time-series
  • Sensor fusion
Code

Sydämen-Nosté Dashboard

AI Accelerator

11/2025 — 02/2026
Web

Backend · scraping · AI extraction

FastAPI + React web app that scrapes company websites, extracts contact information using AI, and stores/manages results in Google Sheets.

outcome → End-to-end lead workflow: scrape → AI extract → review in dashboard → sync to Sheets.

  • FastAPI
  • React
  • Playwright
  • LLM
  • Google Sheets API
Code private

Komee — E-commerce Forecasting

AI Accelerator

10/2025 — 01/2026
Forecasting

Data prep · forecasting · cash-flow features

E-commerce forecasting web app for inventory planning and cash-flow analysis. Used Monte Carlo simulation and Prophet for demand, trend, and seasonality.

outcome → Forecasting and simulation features that support more grounded inventory and cash-flow decisions.

view case study

what i learned → Prophet handles seasonality cleanly, but cash-flow planning needs simulation on top of point forecasts — a single number is rarely useful for a business decision.

  • Python
  • Prophet
  • Monte Carlo
  • pandas
  • Web app
Code private

Video OCR + Audio Report Pipeline

AI Accelerator

01/2026 — 02/2026
OCR

Pipeline design · OCR

Pipeline that turns pipe-inspection videos into structured CSV reports by extracting on-frame text and audio narration.

outcome → Replaced manual report writing with a reviewable CSV export, making the inspection workflow faster and easier to audit.

  • Python
  • OCR
  • Audio processing
  • Video frames
Code private

Jyväskylä Weather & Sensor Dashboard

Lapland UAS · Independent

10/2025 — 12/2025
Sensor

End-to-end (research → implementation)

Combined city LHT and WS100 sensor data with Open-Meteo weather data. Studied precipitation events, humidity, wind, and road-drying behaviour around rain events.

outcome → Interactive dashboard to compare sensor and weather signals before, during, and after rain — used to reason about road conditions transparently.

  • Python
  • pandas
  • Open-Meteo API
  • Dashboarding
Code private

Ravintola Tarantupa Website

02/2026
Web

Independent · full-stack

Bilingual (FI/EN) restaurant website with weekly lunch menu, opening hours, embedded map, and Finland-based time logic for business and lunch hours. Copy / print / share actions for the menu.

outcome → Clear, mobile-first interface where visitors find lunch, prices, hours, phone, and contacts in seconds.

  • React
  • Vite
  • TypeScript
  • i18n
Code
05 — contact

Let's talk.

I'm open to internships, thesis collaborations, and applied data / ML projects — especially around sensor data, forecasting, and practical AI systems. The fastest way to reach me is email.

GitHubLinkedIn+358 451825868
  • Europe/Helsinki
  • Reply time: usually within 1–2 days
  • English (advanced) · Finnish (proficient) · Persian (native)