Mj Pakdel
Mj Pakdel
PROFESSIONAL SUMMARY
A highly skilled and dedicated data science leader, proficient in MLOPs and GenAI with a proven track
record of creating innovative novel solutions and making data-driven decisions for large-scale enterprises.
Collaborative team player with excellent interpersonal skills and ability to effectively communicate
technical information across business units.
ML Techniques: Deep Learning: RNN, LSTM, Autoencoders, Natural Language Processing (NLP),
Large Language Models (GPT-series and BERT-series), RAG, Fine Tuning of LLM, Recommender
Systems, Regression Analysis: Linear and Nonlinear, Classification and Clustering, Time Series Analysis,
Reinforcement Learning, Optimization Techniques, Experimental Design: A/B Testing,
Leadership Skills
Team Leadership: Directed a team of two data scientists, fostering a culture of continuous learning
and guiding them through complex project phases.
Cross-Departmental Integration: Spearheaded the integration of machine learning technologies
across multiple business units, enhancing communication and collaboration between technical and
non-technical teams.
Innovation Leadership: Pioneered the use of advanced machine learning models and MLOps
practices, setting benchmarks for innovation within the organization.
TECHNICAL EXPERIENCE
Programming Languages: Python, Scala, R, SQL
PROFESSIONAL EXPERIENCE
DISNEY STREAMING, Los Angeles, CA Jul 2024 – Present
Staff AI Engineer
Developed an AI-powered chatbot for Onboarding and Internal Support
Generative AI Implementation: Developed, and deployed an AI-powered chatbot leveraging a
Retrieval-Augmented Generation (RAG) architecture to answer employee queries related to IT
systems, HR policies, compliance protocols, and operational guidelines.
Integrated the chatbot with Disney’s internal document repositories, and deployed it using Azure App
Service, managed via Azure DevOps (ADO).
Implemented MLOps best practices, including CI/CD pipelines, environment isolation, and a
monitoring dashboard to track performance, and usage metrics, reducing HR, and IT support tickets
by 35%, and streamlining onboarding processes.
Led the Implementation of Median Mix Modeling (MMM) for Pluto TV:
MLOps Infrastructure Development: Constructed development, staging, and production environments
in Databricks to automate two cloud-agnostic MMM applications: data collection, and model fitting.
Used CI/CD principles for efficiency.
Orchestrated a Kubernetes cluster for a monitoring dashboard, enhancing visual insights into
response curves.
MLOps practices cut retraining times by 90%, allowing weekly updates, and boosting market
responsiveness.
Data Science Implementation: Implemented strategies for managing Pluto TV's $300 million ad
budget, increasing customer acquisition, and sales.
Developed a model using LocalSolver algorithms for MMM parameter fitting.
This enabled continuous optimization of budget across channels, improving executive scenario
analysis.
Achievements included a 4% boost in user acquisition with the existing budget.
EDUCATION
PhD, Economics
Northern Illinois University, Dekalb, IL
CERTIFICATIONS
AWS Certified Machine Learning Specialty, Amazon
AWS Certified Developer-Associate, Amazon
Spark Certified Developer, Databricks