Mangesh (MJ) Bhangare
Director, Enterprise Data Science and AI at Bank of Montreal
Former Director, Data Science and Engineering at Royal Bank of Canada
Data Science and Engineering lead with experience developing large-scale algorithmic systems and data pipelines. Experienced leading teams of Data Science Managers, Data Scientists, Data Engineers and Machine Learning Engineers.
What I bring to the table:
- 10+ years of professional experience solving high-impact data science, AI, ML, & engineering problems.
- 5+ years of experience leading high-performance data teams that focus on implementing breakthrough solutions to support data-driven change and help improve end users' experiences.
- Dedication to building environments that are inclusive, and diverse and respect/leverage individual strengths and differences.
- Experience attracting, developing and retaining top Data Science and Engineering talents.
Former Graduate Student in the Department of Computing Science at Simon Fraser University, Vancouver, BC specializing in Machine Learning and Big Data.
Skillset :Skillset : Language: Python, R, SAS, Java, SQL
Packages/Tech: Apache PySpark, ScikitLearn, Pandas, NumPy, Google Tensorflow, H2O, MLFlow, Palantir Foundry
Domain: Marketing Science, HR/People Analytics, Fraud Analytics, Cyber Security, Climate Analytics, Audit/Compliance, Text Analytics, Customer Analytics and Generative AI.
Modeling experience and primary use cases: Next best Action/Offer, scenario simulation, optimization and recommendation for offers & products, loyalty and frustration, NLP/NLU models, segmentation, new acquisition, upsell, long-term expected value calculation, lift modelling, client journey mapping, price sensitivity, churn modeling and MLOps.
Answering Most Commonly Asked Questions (From my job at RBC and BMO)
Which is your favourite and most commonly used programming language in your team?
-Python
What's your/your team's main Technology Stack?
-PySpark is the main core/computer engine combined with other common Python-based ML libraries
What's your team made of (which profiles)?
-My team has Data Scientists (3 different levels of seniority), Other People's Managers who have their own team of Data Scientists and Machine Learning Engineers
Is your team exploring any GenAI projects?
Quite a few, I am really excited about bringing Gen AI-driven use cases to life in the enterprise, while operating within our full risk appetite.
Does your team only work on POCs or your Machine Learning models are in production?
-It's a combination of POCs on state-of-the-art models and traditional ML models, of which some remain POCs and some get promoted into production depending on the business needs.
What does your current role involve?
-I spend 80% of my time leading different Machine Learning/Data Science/AI projects and 20% of my time is spent on architecting/designing MLOps systems.
Which teams are you currently leading?
- Enterprise AI Services and Capability at BMO - This is a centralized team at the enterprise level, dedicated to supporting a wide range of AI and Data Science initiatives across BMO. The team's primary focus is on creating tangible business value, while also engaging in relevant research efforts.
- Enterprise MLOps and LLMOps at BMO - which is the central capability team dedicated to enhancing overall MLOps practices. The team focuses on implementing the appropriate frameworks, platforms (both cloud-based and on-premises), and reusable libraries to drive efficiency and effectiveness in machine learning operations.
- BMO Gen AI CoP - Central BMO-wide cross-group function which focuses on central capability for the Gen AI initiatives across the bank, LLMOps/Gen AI frameworks, shared models, Model output and evaluation strategies/frameworks, awareness creation, knowledge sharing, etc.
Can I contact you and how?
- Yes, LinkedIn is preferred.
Have More Questions?
- Try this Search
LinkedIn Profile
Git Profile - Grad School/Personal Projects
Professional Experience
CAREER HIGHLIGHTS
Bank of Montreal (BMO), July 2022 to Present, Toronto, Canada
▪ DIRECTOR, DATA SCIENCE AND AI, Enterprise Chief Data Office, July 2022 to Present, BMO
+ I currently lead the Enterprise Data Science and AI team.
At the core, Our team enables the strategic priorities of the bank, by providing progressive data capabilities, coupled with advanced analytical methods to accelerate value generation from actionable insights while operating within the bank’s full risk appetite. We use different state-of-the-art techniques in Computer Vision, NLP, Data Science and AI based on the relevance and need of the business problem.
--------------------------------------------------------------------------------------------------------------------------------------------------------------
Royal Bank of Canada (RBC), 2016 to 2022 (Various Roles), Toronto, Canada
▪ DIRECTOR, DATA SCIENCE AND ENGINEERING, 2021 to 2022, RBC
+ Led efforts to create a client-level enterprise view to identify and define frustration/client sentiment and loyalty using structured and unstructured text data from different sources like web searches, chat-bot interactions, call center transcripts, survey data with NLP/NLU Models and time-series-based discount factors.
+ Led portfolio of different ML models and RBCs ML feature library for product and offer recommendations, contact channel optimization, pricing, profitability and attrition.
+ Co-Led and defined goals for the ML solution engineering team to create an end-to-end data science platform, architecture and data movement across different systems.
+ Developed and implemented different technology transformation projects for Cloud and Hadoop Migration.
+ Engaged Different vendors and technology partners for better planning from a cost and computer perspective to be future-ready.
+Traind data scientists & data engineers and acted as a trusted advocate for machine learning and data science across the enterprise where relevant.▪ SENIOR MANAGER, DATA SCI, 2018 to 2021, RBC
+ Designed and led the development of Daily & Weekly modelling and feature pipeline initiative that enables rapid predictive modelling across different lines of business using technologies such as Spark, Hadoop, Python and Palantir Foundry with Agile Teams and Data Governance Frameworks.
+ Led a team of Data Scientists and Data Engineers to create a C360, central Enterprise-wide repository of client-specific advanced real-time and batch features to support a variety of machine learning models.
+ Developed PySpark-based ML framework, that enables distributed feature preprocessing, model training and distributed scoring for a variety of models trained in different ML libraries, enabling data scientists with different skill levels
to do machine learning at scale in production environments.
+ Also Work as an individual contributor to create marketing propensity, pricing, response, right product/offer fit and channel optimization models for new client acquisition, product upsell and client retention.
+ Conduct high-level reviews to identify client needs and opportunities by analyzing and creating customer use cases, including results, gaps & recommendations to translate them into analytical data applications.▪ DATA SCIENTIST, 2017 to 2018, RBC
+ Developed analytical models for classification, regression, clustering, recommendation systems, text analysis and pricing for product recommendation and offer section.
+ Used Apache Spark and SQL with different database systems to pull data from multiple sources and amalgamate it into new data sets for analytical and modelling purposes and to evaluate key performance metrics and trends.
+ Regularly talked with both business and technical audiences to communicate the business and technical benefits of the proposed analytical solutions.
+ Worked with different lines of business to set up A/B testing to validate the performance of the models in production with Test and Learn Methodology.
+ Worked with the credit bureau data extensively to capture how clients interact with other traditional and non-traditional competition.▪ DATA SCIENTIST ( INTERN ) 2016 to 2016, RBC
+ Part of the RBC Amplify summer internship program.
+ Worked as part of the mortgage predictive modelling team for the innovation
and research department to develop predictive models to better understand the client’s interests, personas and income for the mortgage pre-approval
+ Leveraged scikit-learn and Spark MLlib to create different predictive models
+ Explored deep learning with Tensorflow to make RBC future-ready in the market with deep learning models.--------------------------------------------------------------------------------------------------------------------------------------------------------------
▪ SOFTWARE ENGINEERING - Java and PL-SQL at Accenture PLC, 2013 to 2015
+ Worked as part of the data analytics and performance monitoring team responsible for creating enterprise performance KPIs using technologies such as Oracle PL/SQL and JSP.
+ Worked as a developer to create enterprise resource management (ERM) application using Java as backend with IBM WebSphere server and JSP frontend to manage field technician’s workflow and communication with the supervisors and the central warehouse for a large US Telecom client.--------------------------------------------------------------------------------------------------------------------------------------------------------------
VOLUNTEER POSITIONS - Past and Present
▪ TECHNOLOGY & OPERATION DEI COUNCIL AND WORKING GROUP, BMO
▪ DATA SCIENCE MENTOR, RBC Amplify Program
▪ MENTOR, Toronto Region Immigrant Employment Council (TRIEC)
Education
Simon Fraser University, British Columbia, Canada
Master of Science - MS, Computer Science, Specialization in Machine Learning and Big Data
Project Highlights
Some Relevant Project Highlights some of which I have developed myself and some were created by my talented team. (From my professional career at RBC and BMO)
- A Simulation-Based Credit Card Recommendation Engine, that selects the right product, and right offer as a function of both propensity to respond and expected NIBIT/Profit
- A BERT (Bidirectional Encoder Representations from Transformers) based NLP/NLU models to identify clients' frustration via different unstructured text data points.
- Channel Optimization Model to select the right clients who will respond to Direct Mail
- Call Center Optimization Model to Predict Who is likely to pick up a phone call
- Tele Marketing Model to Predict the likelihood someone will say yes to a specific request made over a phone call
- Marketing Model to Predict who is likely to Open a specific account in the next X-Days (X: Number of Days)
- Anomaly Detection Model to predict a rare event based on Gaussian Mixture Model
- AutoEncoder Based Anomaly Detection/Rare Event Prediction Model
- AutoEncoder-Based Feature Reduction
- A Model for Small Businesses to help them select the right product
- And many more, but unfortunately can't talk about it here due to the nature of the project
- https://about.bmo.com/ai-turbo-charges-bmos-drive-to-data-automation/
Please note due to confidentiality agreements, I can't disclose all projects and their details that I am working on. If you find something interesting on my profile please reach out on Linkedin, I may be able to share high-level ideas or concepts that are not bonded by NDAs.
Volunteer Experience
Why Mentor? (Taken from triec website) Mentors are leaders. They give back to their professional communities. They coach people to achieve their full potential and work effectively with people of all different backgrounds and experiences in today’s diverse workplace. Giving just 18 hours of your time over three months could help change someone’s life.
My YouTube Videos
Please be aware that I am not a content creator on YouTube, and I do not regularly upload or update videos. However, feel free to take a look and enjoy!
LET'S CHAT
Thanks for Visting hoping to connect soon
Copyright Ⓒ 2022 Mangesh B