Open to opportunities

Shakiba Tasharrofi

@shakibatasharrofi

Message

Senior ML Engineer building end-to-end production AI systems across anomaly detection, payments, and GenAI.

Netherlands

Message

What I'm looking for

I’m looking for a role where I can own end-to-end ML systems—deployment, monitoring, drift detection, and retraining—while integrating GenAI responsibly to deliver measurable business value with production-grade engineering.

I’m a Senior ML Engineer with 4+ years of experience in financial technology, specializing in infra fraud detection, payments systems, and production ML deployment. I own end-to-end ML lifecycles—from conceptual design through deployment and monitoring—using graph neural networks, anomaly detection, and GenAI integration.

At Mithra-ai Solutions, I independently architected and deployed a multi-agent AI system using Google Agent Development Kit (ADK) and LangGraph, taking full ownership from concept to production. I built and maintained Document Intelligence processing 30K+ invoices quarterly with end-to-end monitoring and retraining pipelines on GCP (Vertex AI, Cloud Run), and integrated GenAI with Gemini API/LLMs to implement confidence scoring, drift detection, and automated feedback loops.

Previously at Bale Financial Messenger, I drove recommender systems and multi-output regression models into production on Kubernetes, built monitoring metrics (CTR, engagement rates, latency), and improved business outcomes through systematic A/B testing. I’ve also developed large-scale churn prediction and text de-duplication solutions, and I enjoy collaborating cross-functionally to deliver practical, production-ready ML.

Experience

Work history, roles, and key accomplishments

Current

ML Engineer

Current

Mithra-ai Solutions

Dec 2023 - Present (2 years 7 months)

Architected and deployed a multi-agent AI system in production using Google Agent Development Kit (ADK) and LangGraph, taking full ownership of the project lifecycle. Built an invoice document intelligence pipeline and integrated Gemini-based GenAI with monitoring, drift detection, retraining pipelines, and production-ready deployments on GCP.

Agent Development Kit (ADK)LangGraph Document Intelligence Invoice Processing Vertex AI Cloud Run GCP Gemini API Monitoring Drift Monitoring Retraining Pipelines Docker CI CD Python Django Model Monitoring

Data Scientist and ML Engineer

Bale Financial Messenger

Sep 2021 - Dec 2023 (2 years 3 months)

Developed and deployed recommender and regression ML systems to production, including real-time personalized financial reporting for 200,000+ users. Improved engagement metrics via A/B testing, automated retraining based on performance degradation, and NLP pipeline optimization.

TensorFlow PyTorch Kubernetes A B Testing CTR Optimization XGBoost scikit learn Multi Output Regression Model Monitoring NLP Feature Engineering Recommendation Systems

Data Scientist

Yaas Arghavani System Engineering

Apr 2021 - Sep 2021 (5 months)

Built and deployed churn prediction for Mellat Bank’s mobile banking application using Random Forest and XGBoost, including feature engineering and financial-services evaluation metrics. Developed a time-series prediction system with Keras and TensorFlow aligned to the bank’s technology stack and compliance requirements.

Random Forest XGBoost Feature Engineering Evaluation Time Series Keras TensorFlow

Data Scientist

Institute of Applied Intelligent Systems of Tehran Universit

Jan 2020 - Sep 2020 (8 months)

Developed and deployed a text de-duplication engine for large-scale complaint processing using Locality-Sensitive Hashing, with efficient pipelines and monitoring. Designed multi-label/multi-class complaint segmentation and implemented monitoring metrics including accuracy, precision-recall curves, and data drift detection.

Locality Sensitive Hashing Data Pipelines Monitoring Precision Recall Curve Monitoring Drift COVID 19 Surge Robustness

AI Researcher

Digikala Next

Aug 2019 - Nov 2019 (3 months)

Conducted R&D on fraud detection in e-commerce payments using deep learning and graph neural networks. Explored anomaly detection techniques and validated approaches via proof-of-concept development using real transaction data.

Fraud Detection E commerce Payments Deep Learning Anomaly Detection Proof of Concept

Teaching Assistant, Advanced Programming

Amirkabir University

Jan 2019 - Jun 2019 (5 months)

Instructed students in Python and C++ programming fundamentals and provided technical guidance on code quality, debugging, and software development best practices. Developed teaching materials and mentored students on practical programming skills.

Python C++Teaching Mentorship Code Quality Debugging Software Best Practices

Education

Degrees, certifications, and relevant coursework

Amirkabir University of Technology

Master of Science, Digital Electronics Systems

2018 - 2020

Activities and societies: Ranked 3rd among Master’s participants. Thesis focused on Graph Neural Network applications for financial fraud detection.

M.Sc. in Digital Electronics Systems with a thesis on designing and implementing a graph convolutional network for fraud detection in financial systems. Completed relevant coursework including Statistical Learning, Pattern Recognition, Big Data, and Signal Processing.

Amirkabir University of Technology

Bachelor of Science, Electrical, Electronics and Communication Engineering

2014 - 2018

Activities and societies: Honored as “Exceptional Talents,” awarded a fellowship, and received an international internship in China.

B.Sc. in Electrical, Electronics and Communication Engineering. Completed the degree with additional recognition including awards and an international internship.