VAHE HAMBARDZUMYAN // AI_ML

// QUEST LOG //

EDUCATION & EXPERIENCE

🎓

2021EDUCATION

B.Sc. Computer Science

▸ French University in Armenia · 2021 – 2027

Relevant coursework: Data Structures, Algorithms, Database Systems, AI, Machine Learning, Software Engineering, Calculus, Linear Algebra, Statistics. Collaborated on bilingual French-English group projects strengthening communication and problem-solving skills.

1200 Education XP

🎓

2024EDUCATION

B.Sc. Computer Science

▸ University of Toulouse · 2024 – 2027

International exchange programme deepening theoretical CS and AI foundations. Strengthened cross-cultural collaboration and technical communication in French and English academic environments.

900 Education XP

🧠

2024TRAINING

ML / DS Student

▸ Picsart Academy · 2024 – 2025

Advanced ML/DS programme covering Deep Learning (CNNs, RNNs, Transformers), NLP, Computer Vision, Reinforcement Learning, and Generative AI including diffusion models. Strong mathematical grounding in Calculus, Linear Algebra, Probability, and Optimization.

2200 Training XP

📊

2025WORK · CURRENT

Data Analyst

▸ Trendsetter A Multibrand · Sep 2025 – Present

Hands-on end-to-end ETL processes ensuring clean, reliable data pipelines for analytical use. Business intelligence workflows supporting reporting and strategic decision-making. Interpreted complex datasets and presented insights through interactive Power BI and Tableau dashboards.

1800 Work XP · ONGOING ▮

// CHARACTER STATS //

TECH SKILLS

ML / Deep Learning

Python95

scikit-learn88

Neural Networks / CNNs / RNNs82

Transformers (GPT/BERT/CLIP)80

Reinforcement Learning78

Data & Analytics

SQL / DBMS85

ETL Pipelines82

Apache Airflow75

Power BI / Tableau83

A/B Testing72

AI Specializations

Computer Vision82

NLP / LLMs80

Generative AI (DDPM/DDIM)76

Mathematics (Calculus/LinAlg/Stats)90

OOP / SOLID / Algorithms85

// INVENTORY //

PROJECT ARTIFACTS

◆ EPIC

ITEM_001 · REINFORCEMENT LEARNING

BANDIT PHARMA AI

Multi-Armed Bandit algorithms for personalized drug-dosage optimization. Implemented ε-greedy, UCB1, and Thompson Sampling from scratch across 3 patient groups and 15 drug-dosage arms. Thompson Sampling cut regret by 12–27% vs baselines across 100 Monte Carlo runs.

PYTHONNUMPYSCIPYMATPLOTLIBCOLAB

15 arms

100 MC runs

-27% regret

⌥ VIEW ON GITHUB →

◈ RARE

ITEM_002 · NLP / DATA SCIENCE

ARMENIAN NLP ENGINE

Armenian news website scraper and NLP analysis pipeline targeting sentiment, tone, and media bias in political and economic coverage. Built to improve Armenian language datasets for LLM training — a language severely underrepresented in modern AI corpora.

PYTHONBEAUTIFULSOUPSELENIUMNLPPANDAS

NLP pipeline

ARM language

LLM dataset

⌥ VIEW ON GITHUB →

· UNCOMMON

ITEM_003 · CLASSIFICATION

PERSONALITY CLASSIFIER

Introvert vs Extrovert binary classifier using multiple ML algorithms. Full pipeline from data preprocessing and feature scaling to model training and evaluation. Benchmarks MLP, SVM, and Random Forest classifiers with confusion matrix and classification report output.

SKLEARNMLPSVMRANDOM FORESTMATPLOTLIB

3 models

F1 scored

binary clf

⌥ VIEW ON GITHUB →

// ABILITY TREE //

WHAT I BUILD

MAX LVL🧠

DEEP LEARNING

CNNs, RNNs, Transformers. Computer vision, NLP, and multi-modal architectures. Strong mathematical grounding in optimization and probability theory.

LVL 8🔬

GENERATIVE AI

Diffusion model fundamentals including DDPM and DDIM. Understanding of latent space manipulation and generative architecture design.

LVL 8🎮

REINFORCEMENT LEARNING

Multi-armed bandits, policy optimization, reward shaping. Applied RL to real-world medical optimization with measurable regret reduction.

LVL 9📊

ANALYTICS & BI

End-to-end ETL pipelines, SQL, Apache Airflow, Power BI, and Tableau. Translating raw data into actionable business insights.

LVL 8💬

NLP & LLMs

GPT, BERT, CLIP architectures. Transformer fine-tuning, Armenian NLP dataset work, sentiment analysis and media bias detection.

LVL 7🌍

MULTILINGUAL

Armenian (native), French (fluent), English (fluent), Russian (fluent), Spanish (elementary). Collaboration across cultures and academic institutions.

VAHE HAMBARDZUMYAN

EDUCATION & EXPERIENCE

TECH SKILLS

PROJECT ARTIFACTS

WHAT I BUILD

GET IN TOUCH