Available for Opportunities

Michael Hakkinen

AI Engineer  ·  Full-Stack Developer  ·  NLP Researcher

Machine Learning NLP / DeBERTa Python · JavaScript Flask · Laravel PyTorch · TensorFlow
2+ Years Experience
7+ Projects Built
2 IEEE Papers
Michael Hakkinen
Michael Hakkinen
AI & Full-Stack Dev
Python / ML
NLP / BERT
JavaScript
Laravel / PHP
Open to Work
Scroll
Michael Hakkinen

Michael Hakkinen

Artificial Intelligence & Full-Stack Developer

Work Experience

2026

IT Operations & Developer (Remote)

Anradus Pte Ltd  ·  Singapore
  • Supported design, development, deployment, and maintenance of internal business systems and APIs
  • Troubleshot application, database, and system issues to minimize operational disruptions
  • Implemented software improvements and bug fixes using JavaScript and SQL
  • Participated in code reviews and improved coding standards across the team
  • Leveraged AI-assisted development tools to accelerate productivity and code quality
TypeScript PostgreSQL Github IT Operations QA & Testing
2024 — 2025

Artist Hospitality Coordinator

Color Asia Live  ·  Jakarta, Indonesia
  • Delivered hospitality operations for major concerts with 5,000 – 20,000+ attendees, coordinating backstage logistics end-to-end
  • Achieved 100% on-schedule execution rate across all managed events through rigorous planning
  • Directly supported artists including Mariah Carey, Jonas Brothers, Josh Groban & David Foster — maintaining 100% positive feedback
Event Management Stakeholder Relations High-Pressure Operations
2022 — 2023

Application Developer (Research)

Bina Nusantara University  ·  Jakarta, Indonesia
  • Collaborated with a team of 3 developers and 1 database specialist to implement and evaluate BERT-based models, improving text classification accuracy by up to 12% compared to baseline models.
  • Applied SMART fine-tuning techniques across multiple BERT variants, increasing generalization performance by 3%10% on unseen datasets.
  • Designed and executed experiments comparing model accuracy, loss trends, and training efficiency, reducing experiment runtime by 20% through optimized preprocessing pipelines.
  • Extracted raw text documents into a database using Cheerio + Node.js, automating 100% of dataset ingestion and reducing manual processing time by 90%.
  • Authored research documentation and evaluation reports, improving team research clarity and collaboration by 70%.
  • Co-authored a research paper contributing to the academic community's understanding of NLP advancements.
BERT Model Natural Language Processing Python Keras TensorFlow

Education

2019 — 2026

Bachelor of Computer Science

Bina Nusantara University  ·  Jakarta, Indonesia
  • Coursework: Software Engineering, Machine Learning, AI, Algorithm Design, Data Structures — consistently scoring 85%+
  • Built multiple full-stack and ML systems improving model accuracy by 20–35% through fine-tuning and pipeline optimization
  • Conducted NLP experiments with BERT & DeBERTa — achieved 10–15% accuracy improvements over baselines
  • Designed preprocessing pipelines cutting processing time by 30%; co-authored research published in IEEE
Python · Java · C++ · SQL PyTorch · TensorFlow · Keras BERT · DeBERTa · NLP Data Pipelines
2007 — 2019

Primary — High School

Universal School  ·  Jakarta, Indonesia

Academic & Leadership

  • Consistently ranked top 10 in class; scored 80%+ in Mathematics, English, Computer Studies
  • Served as class representative; coordinated school events and inter-party communication
  • Achieved consistent top placements in inter-school badminton & basketball tournaments
Portfolio

Selected Work

ML research, full-stack systems & NLP

View GitHub

Other Projects & Publications

Pokedex App
Web · Vue.js
Pokedex

Interactive Pokédex built with Vue.js, featuring search and pagination

View Live
Coalmade Website
Web · HTML/CSS/JS
Coalmade — Company Website

Corporate website for a coal trading company

View Live
BERT SMART Research Paper
Research · IEEE
BERT vs SMART Fine-Tuning

Comparative NLP research paper

Read Paper
E-commerce Sentiment Analysis
Published · IEEE
E-Commerce Sentiment — IEEE

Lexicon-based analysis · Published in IEEE

View Publication
BERT Research Paper
Research · NLP
BERT Fine-Tuning Research

In-depth fine-tuning analysis for classification

Read Paper

Let's Work Together

Whether you have a project in mind, a role to fill, or just want to connect — I'm always open to a conversation.