Mailén Gómez Mayol Email & Phone Number

Buenos Aires Province, Argentina

Buenos Aires

Buenos Aires

- Deploy and maintain high performance machine learning models for optimizing mobile games monetizing through ads.- Collaborate with cross-functional teams, including data analysts and software engineers, to integrate machine learning models into existing systems.- Actively participate in code reviews, architectural discussions, and technical documentation to promote best practices and knowledge sharing within the team.- Generate regular reports and dashboards to communicate performance trends, revenue metrics, and optimization results to stakeholders and senior management.- Conduct regular A/B testing and performance analysis to iterate and refine ad waterfall configurations, improving fill rates, eCPM, and overall ad revenue

Buenos Aires, Argentina

- Contributed to the establishment and training of the machine learning team within the AdTech team.- Contributed to developing innovative machine-learning models to optimize ad revenue.- Implemented robust monitoring systems to ensure the optimization model data and infrastructure's stability, performance, and availability.

Buenos Aires, Argentina

As a quantitative analyst at Crisil, I work within the Data Science and Artificial Intelligence team of a tier 1 bank as a model validator.I validated more than 15 models based on machine learning techniques (logistic regressions, decision trees, NLP, neural networks).I worked with models of :- Fraud detection in debit and credit cards, transactions, loans, and identity.- Electronic communications surveillance.- Transcription and classification of speech in English and Spanish.Part of the validation work consists of developing simple benchmark models, and the calculation of different performance metrics (KPI).Research project: NLP based automatic ESG reporting. This project involves parsing PDF files and question answering models. I used an algorithm based on a LSTM autoencoder to parse PDF files into plain text. I explored other computer vision techniques like convolutional neural networks.For the question answering section of this project I explored different NLP methods such as knowledge graphs and transformers (different implementations of BERT: albert, BERT+SQuAD)

Buenos Aires, Argentina

I worked on the design and development of the QC-radar algorithm for weather radar data quality control. The algorithm consists of a series of filters that clean and improve the weather radar data according to the use specified by the user: display of weather radar data images on the web page, data assimilation for the numerical forecast model, precipitation estimation, among others.The scientific development of the model was in charge of Dr. Juan Ruiz et al, mainly in Fortran for the optimization of computation time. My contribution to the algorithm was the development of the schema for the production implementation entirely in Python for flexibility.I fully documented the algorithm and all its filters.

Buenos Aires, Argentina

As an intern in the R&D department I generated products of meteorological interest using neural networks.- I designed and developed a model based on convolutional neural networks, which generates synthetic weather radar images from lightning geolocation data. Data from weather radar are indispensable for severe weather monitoring. The model I developed is useful for reconstructing missing parts and filtering out possible noise in weather radar data.I developed the model using TensorFlow and Python. I analyzed the data used to train and validate the model using specific Python libraries for weather radar data: pyART, as well as, pandas, numpy.- I designed and developed an lightning parameterization model using lightning location measurements and numerical simulations. The model is based on convolutional networks.I developed the model using Python and pytorch.The training data were reanalysis of the CFS model and lightning location measurements from the Vaisala GLD360 global network. The validation data were simulations of WRF numerical prediction model. I analyzed the data using Python, pandas, numpy and wrf-python.

Master'S Degree, Statistics

Specialist, Mathematical Statistics

Msc Physics, Machine Learning, Modeling Of Physical Processes In The Atmosphere, Remote Sensing