Open to opportunities

Vahram Babadjanyan

@vahrambabadjanyan

Message

I am an Applied Data Scientist and ML Engineer delivering production-grade models for finance and vision.

Armenia

Message

What I'm looking for

I seek an end-to-end ML role where I can own model development and productionization—especially in finance, risk, computer vision or NLP—within a data-driven, collaborative team focused on business impact.

I am an applied Data Scientist and Machine Learning Engineer with four years of industry experience in financial services and research-driven computer vision. I focus on delivering production-grade models for credit default prediction, fraud detection and risk modeling.

I have led end-to-end ML projects, from SQL-based feature extraction and model training to deployment, monitoring and reporting. My work has driven measurable business impact, including reductions in default loan rates and production deployments for banking and defense applications; I also open-sourced Deeva, an object detection dataset explorer.

I take ownership of data pipelines, feature engineering, model fine-tuning (including transformer models) and deployment automation. I am comfortable working with modern ML stacks and tooling to ship, monitor and maintain reliable models in production.

Experience

Work history, roles, and key accomplishments

Machine Learning Engineer

Ministry of Defense

Jul 2023 - Jul 2025 (2 years)

Served as a machine learning engineer in the AI division, training and deploying object detection and segmentation models (YOLOv11, Detectron2) for mission-critical visual recognition systems and integrating custom data augmentation into Ultralytics pipelines. Led dataset curation and annotation correction, developed the Deeva dataset explorer, and created automated web parsers for NLP preprocessi

OpenCV PyTorch Detectron2 YOLOv11 Streamlit Apache Airflow NLP

Deeva — Dataset Explorer

Deeva

Mar 2024 - Jan 2025 (10 months)

Built Deeva, a Streamlit-based tool for fast, intuitive exploration, versioning, and reporting of object detection datasets to support dataset management and model development workflows.

Streamlit OpenCV PyTorch Data Visualization Python

Data Scientist

EvocaBank

Jan 2023 - Jul 2023 (6 months)

Designed, developed, and deployed a full-cycle ML pipeline for loan default prediction on EvocaBank’s core retail product ActionLoan, covering SQL-based feature extraction through production deployment. Built regression models for macroeconomic forecasting used in central bank policy simulations and created a Streamlit monitoring dashboard for real-time model tracking.

SQL Streamlit Plotly scikit learn Pandas XGBoost LightGBM fastAPI

Autopt — HPO Toolkit

Autopt

Oct 2022 - Jul 2023 (9 months)

Developed Autopt, a Python package for automated hyperparameter tuning using BayesianSearchCV with cross-validation utilities and top-N estimator selection for regression and binary classification.

Python scikit learn Bayesian Optimization Cross Validation Packaging

Machine Learning Research Engineer

FAST Foundation/8nook

Jul 2021 - Oct 2022 (1 year 3 months)

Initiated and delivered default loan detection pipelines for two Armenian banks, contributing to 23% and 35% reductions in default rates by integrating behavioral risk profiles and dynamic credit scoring. Researched clustering in high-dimensional feature spaces and implemented XML/PDF extraction tools to integrate unstructured financial documents into ML pipelines.

Python SQL scikit learn XGBoost CatBoost Clustering Lxml

Data Scientist

IDBank

Apr 2021 - Jul 2021 (3 months)

Automated cross-departmental data exchange workflows using MS SQL Server and pyodbc, improving data flow efficiency and accuracy. Designed a customer evaluation system based on financial behavior analytics and automated scheduled reporting processes to reduce manual effort.

MS SQL Server Python JupyterLab Reporting

Data Analyst

Rosgosstrakh Armenia

Jan 2020 - Apr 2020 (3 months)

Analyzed cross-sales data to identify trends and produced standardized CRM report templates to improve stakeholder communication. Performed comprehensive data cleaning and preprocessing to support sales strategy optimization.

Excel Pivot Tables Data Cleaning