Matias Leone Email & Phone Number

New York City Metropolitan Area

New York, Us

Ads recommender models: ranking, retrieval, embeddingsMLOps: Feature Store, model serving platform, training pipeline, vector DBTensorFlow, AWS, Python, Golang

Menlo Park, Ca, Us

AI development infrastructure team. Building internal AI tools fo ML engineers, to allow them train the next-gen of ranking models for Ads and Feeds, using PyTorch and distributed training on GPUs.

Chicago, Illinois, Us

eCommerce - search & recommendationsClosing the gap between data scientist and backend engineers. Design, lead and implementation from scratch of the ML inference platform used to perform model predictions at scale across the company, supporting model versioning, deployment, feature store and monitoring. Using TensorFlow, AWS, Java, Python, Annoy, Cassandra, Memcached, S3. Lead the re-architecture of the e-commerce search engine to transition to a fully ML-driven approach, using a 2-step recommender design with a candidate selection model and a personalized ranking layer.Design, lead and implementation of the an A/B test platform to perform online ML experimentation.

San Francisco, Ca, Us

Leading a 10+ people development team on multiple AI and data pipelines initiatives for different clients, taking on a role of lead machine learning engineer and data architect:- Real-time analytics: modernization of legacy on-premise data warehouse into a real-time data pipeline on Azure using Kafka, Spark streaming and Snowflake. Platform design from scratch to ease data access by quant team to perform ML.- Predict assets volatility with ML: development of a machine learning model for the prime brokerage business of a top US investment bank to predict volatility of their client assets and make margin loan decisions. Using TensorFlow to train a deep recurrent neural networks and perform time-series analysis, Python, Jupyter notebooks, Scala, feature engineering, regression, trained on GPU with 3+ years of historical data. Design, implementation and deployment to production end-to-end, deployed as a Python REST service to be consumed by other systems.- Demand coverage engine: development of a trading algorithm for a top US investment bank, to match supply and demand of financial assets in the best possible way. Solving a 1M equations combinatorial problem with linear programming, Gurobi, Scala, Spark, Hadoop. Mathematical model design, implemented and deployed to production end-to-end.- Data-lake platform for analytics: design and development of a centralized on-premise data-lake for analytics for a top investment bank. Using Hadoop, Spark, Scala, Java, Python, HBase, HDFS, Impala, Kudu, Oozie, Spring Boots, Cloudera. Platform was designed and implemented from scratch enabling multiple teams to develop analytics projects on top of it.- Data reconciliation with AI: development of a machine learning model for a hedge fund IT provider to compare and reconcile multiple data sources with different securities descriptors. Training a neural network with TensorFlow, Python, Jupyter notebooks. End-to-end implementation and integration with existing C# backend.

San Francisco, Ca, Us

Leading a team to design and develop a Big Data processing hub and data warehouse architecture that process two hundred million transactions per month coming from one hundred treasury service systems, in a near real-time fashion. Including solutions to handle disaster recovery, data archival and multihoming deployments.- Using Java 7 and Big Data technologies: Apache Hadoop, MapReduce, HBase, Hive, Pig, Cloudera distribution, Apache Storm, Apache Kafka, Apache Flume and ZooKeeper.- Coordinating actions with others development teams in New York and India.

Buenos Aires, Buenos Aires, Ar

Professor of Real Time Computer Graphics course at the National University of Technology (UTN-FRBA) in Argentina- Course web site: http://tgcutn.com.ar/- 3D graphics engine developed for this course: https://github.com/leonematias/TgcViewer- Topics: low level understanding of real time rendering, graphics pipeline, 3D transformations, animation, lighting, shaders, DirectX.

Buenos Aires, Buenos Aires, Ar

Adjunct professor of Database design course at the National University of Technology (UTN-FRBA) in Argentina.Topics: databases internal physical design, SQL language, model design patterns, indexes and constraints, B-Tree, usage of Oracle and SQL Server

Af

Leading a team of eight software engineers in different projects and technologies to develop web portals and applications for the public education industry.- Development of multiple web applications in Java, PHP and PostgreSql cluster using a Big Data modeling approach.- Architecture design and implementation of a centralized content management system (CMS) to handle all organization media assets.- Designing the software architecture, data architecture, service architecture and construction of the company framework.- Data Warehouse modeling for a decision-making system.- Coordination of internal teams and external vendors.

Leading a team of three developers to implement a unified web application for the mainpublic health insurance of Argentina.Design and development of internal core web framework that was then leveraged by multiple other projects.Implementation of deployment of web site end-to-end.Using Java, PHP, Oracle, HTML, Javascript, Eclipse, SVN, Maven.Performance tuning of Oracle database queries.

Design and development of internal core framework used to develop more than ten different web applications by other teams.Using Java, Oracle, Google Web Toolkit, HTML, Javascript, Spring, Hibernate, Maven, SVN.Development of multiple core UI widgets.

Web development to improve capabilities of e-learning applicationUsing Java, Oracle, HTML, Javascript, Spring, Hibernate

Master'S Degree, Software Engineering

Bachelor Of Science (Bs), System Engineer