Subhargh Mazumder

AI Researcher

Singapore, Singapore11 yrs 10 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in building large scale machine learning applications.
Led cross-platform recommendation systems for millions of users.
Strong background in data-driven product management.

Stackforce AI infers this person is a Data Science and Machine Learning expert in the Media and Fintech industries.

Contact

Skills

Core Skills

Machine LearningData Science

Other Skills

AlgorithmsProduct ManagementPythonScalaMySQLData EngineeringData AnalysisStatisticsData MiningProgrammingBusiness StrategyBusiness AnalysisBusiness IntelligenceBusiness AnalyticsAnalytics

About

AI, Data Science, Distributed Systems and Products specialist, experienced in building large scale machine learning applications, data-driven product management & analytics. Languages & Platforms: Scala, Spark, Python, R, MySQL, C, C++ Techniques: Linear Regression, Logistic Regression, Bagging, Boosting, Random Forests, Gradient Boosted Trees, Deep Neural Nets (DNN), Convolutional Neural Nets (CNN), Recurrent Neural Nets (RNN), Natural Language Processing (NLP), Naive Bayes, Bayes Nets, Bayesian Optimization, Probabilistic Graphical Models, Clustering, Spectral Methods, Time Series Analysis Technologies: MySQL, PostgreSQL, Spark, Hadoop, HDFS, Redis, MongoDB, DynamoDB, Cassandra, Amazon Redshift, AWS Lambda, Amazon Kinesis, Apache Kafka

Experience

Twitter

Senior Machine Learning Engineer

Feb 2022 – Present · 4 yrs 1 mo · Singapore

Ex Senior Engineer in Relevance Platform team which works on user recommendations building blocks (candidate generation, similarity, embeddings) to serve product surfaces like Home Timeline, Notifications, Email and Who-To-Follow.

Machine LearningData ScienceAlgorithmsProduct ManagementPython

Mediacorp pte ltd

Assistant Vice President

Sep 2020 – Jan 2022 · 1 yr 4 mos · Singapore

Technical Lead & Architect of cross platform recommendations across all MediaCorp properties
Leading a team of data scientists, machine learning and data/software engineers building out the entire backend infrastructure for
data collection, processing, and machine learning model development cum real-time inference for serving recommendations at
scale across all MediaCorp websites (Channel News Asia, TodayOnline, meWatch, meListen, 8world, 8days & others)
We serve article, video (movies, series, news & short clips) and audio (music & podcast) recommendations across 20 million +
monthly active users spanning all properties focussing on Singapore and Southeast Asia region
Transitioned from a siloed site-specific data and ML models to one considering user interests across all sites and formats
Created unified embedding for disparate content types (article, audio, and video) and users based on incremental weighted avg
Productionised and experimenting with different models starting with content based and collaborative filtering to session based,
including Deep Factorisation Machines and DCN v2 (Deep & Cross Network) for ranking
Setup the entire monitoring, retraining and alerting pipelines with the data engineering team to automate model deployments
New models achieved more than 2x CTR lift on meWatch (video content) and TodayOnline, up to 0.5x on CNA and 8world
Tools: Python, Docker, Kubernetes real time model serving/training, Scala/Spark for batch/streaming data pipelines, Redis, MySQL

Machine LearningData SciencePythonScalaMySQL

Tookitaki

Machine Learning Research & Engineering Lead

Jan 2019 – Aug 2020 · 1 yr 7 mos · Singapore

Leading an 8 member team of Research Data Scientists, Scala Functional Programmers and Data Engineers to build the next generation of Spark & AI based reconciliation product.

Machine LearningScalaData Engineering

Vymo

Lead Data Scientist

Feb 2017 – Jan 2019 · 1 yr 11 mos · Bengaluru Area, India

Vymo is a Mobile-first AI driven Sales Assistant.
Product Owner for "Suggestions", Vymo's Contextual Recommendation Engine for real-time interventions. More details: http://getvymo.com/ai

Machine LearningData ScienceProduct Management

Dream11.com

Senior Data Scientist

Aug 2015 – Nov 2016 · 1 yr 3 mos · Mumbai Area, India

Dream11.com is the world's largest Fantasy Cricket & Football Game, with an exponentially growing user base of 1 million+ users. Our mission is to make every game exciting for sports fans globally. We are a mobile-first Series D funded startup in the sports, gaming and e-commerce sector.
As a part of the Products and Data Science Teams, I handle everything data @ Dream11, from sourcing to science.

Data ScienceData AnalysisMachine Learning