Looking for a job

Manuk Manukyan

@manukmanukyan

Message

I build AI, RAG, fraud detection, and data pipelines that turn complex data into decisions.

Armenia

Message

What I'm looking for

I'm looking for AI/ML, data science, or data engineering work where I can build production-minded models, RAG systems, fraud detection, and data pipelines while continuing to grow through challenging applied problems.

I've built AI and data systems at DigiTrader, Evoca Bank, and MathrixAI, from photo-authenticity and fraud-detection pipelines to SQL reporting datasets and RAG workflows.

At DigiTrader, I developed perceptual-hashing and pretrained classifier workflows to flag recycled and AI-generated listing photos, alongside rule-based fraud detection and a constrained delivery-matching engine.

I've also evaluated CNN robustness across 25 corruption conditions, analyzed GPT-2 induction heads, and built customer-segmentation dashboards using Python, SQL, PyTorch, Airflow, Streamlit, and Plotly.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

DigiTrader

May 2025 - Present (1 year 4 months)

Built a photo-authenticity pipeline using perceptual hashing and a pretrained deepfake/AI-image classifier to flag AI-generated listing photos. Developed rule-based fraud detection for escrow and dispute resolution, and designed a detour-miles delivery-matching engine as a constrained optimization and ranking problem.

Python PHash Deepfake Detection CV Classification Rule Based Fraud Controls Constraint Optimization Ranking

AI Research Intern

MathrixAI

Feb 2026 - May 2026 (3 months)

Built and evaluated RAG pipelines combining embedding models, vector retrieval, and LLM integration. Authored technical analyses on optimization algorithms, backpropagation, regularization, and loss function design.

RAG Embeddings Vector Retrieval LLM Optimization Algorithms Backpropagation Regularisation Loss Functions

Data Engineering Intern

Evoca Bank

Sep 2025 - Dec 2025 (3 months)

Designed SQL-based ETL pipelines with Airflow orchestration, retry logic, and pre-load validation, delivering reporting datasets across 3 business units. Built a K-Means customer segmentation system for 5,000 clients with a Streamlit and Plotly dashboard, validated using ANOVA, Chi-Square, and Welch's T-Test.

SQL PostgreSQL Airflow K Means Streamlit Plotly Scipy ETL Pipelines Data Validation

Data Analytics Engineering Trainee

EPAM Systems

Feb 2025 - May 2025 (3 months)

Completed advanced PostgreSQL training covering CTEs, window functions, subqueries, query optimization, and DCL/DDL/DML. Modeled real-world schemas for a Real Estate Agency and Subway system.