Santiago Hernández, Developer in Buenos Aires, Argentina
Santiago is available for hire
Hire Santiago

Santiago Hernández

Verified Expert  in Engineering

Machine Learning Engineer and Developer

Location
Buenos Aires, Argentina
Toptal Member Since
January 29, 2019

Santiago is a machine learning engineer keen on solving large-scale data, optimization, and engineering problems. 他带领数据科学家团队构建了预测全球超过3亿用户价值的产品,以优化世界领先公司的广告. Santiago considers it a serious undertaking when developing elegant, efficient, and well-designed software systems.

Portfolio

ZeeMaps (via Toptal)
Bootstrap, PostgreSQL, Redis, Plotly, IPython Notebook, Docker, SaaS, Dash...
MercadoLibre (with Mutt Data)
Docker, Machine Learning, IPython Notebook, Optimization, Programming...
Jampp
Pandas, Scikit-learn, Tornado, C, Presto, PostgreSQL, Bash, Git...

Experience

Availability

Part-time

Preferred Environment

Bash, Vim Text Editor, Python, Git, Linux

The most amazing...

...我开发了一个机器学习系统,每小时可以传输超过6000万条消息,并实时构建数据集来预测移动应用的盈利情况.

Work Experience

Data Scientist

2019 - PRESENT
ZeeMaps (via Toptal)
  • 分析公司的系统即服务(SaaS)数据,以找到见解并制定战略业务决策.
  • 建立一个仪表板来评估公司的发展,方便地看到分析和辅助决策.
  • 将订阅用户分为获得用户和风险用户,使我们能够专注于防止风险用户流失, driving the company's growth.
  • 提供收益、终身价值、流失率和其他用户类别指标的可见性.
  • Tagged users in high and medium churn risk to allow focused support and customer service.
  • Analyzed and optimized sponsored search campaigns to acquire new users.
  • Dockerized the setup and deployment of the dashboard system.
  • 执行每个计划和每个用户的成本分析,以评估当前SaaS的定价方法,并在适当的时候通过与高客户执行交易来提高收入.
Technologies: Bootstrap, PostgreSQL, Redis, Plotly, IPython Notebook, Docker, SaaS, Dash, Python, Data Science

Machine Learning and Optimization Leader

2019 - 2020
MercadoLibre (with Mutt Data)
  • 定义了一个优化问题,以优化营销团队的预算和目标ROAS(广告支出回报),目标分配来自Google Shopping的收入.
  • 对公司和营销团队的数据进行探索性数据分析,以了解预算之间的关系, goals, and results.
  • Developed a system to forecast and predict cost, 谷歌购物活动的收入和投资回报取决于预算和目标ROAS目标受业务约束.
  • 设计并开发了一个系统来解决优化问题,并根据预测确定如何在Google Shopping上设置数字营销活动, forecasts and business constraints.
  • Dockerized the setup and deployment of the system using docker containers and docker compose.
  • Deployed the system for two accounts in a single country, after the results, on five more accounts on the same country, followed by rolling out to other countries.
Technologies: Docker, Machine Learning, IPython Notebook, Optimization, Programming, Prophet ERP, Pandas, Scikit-learn, Forecasting, Data Science, Python

Machine Learning Engineering Leader

2015 - 2019
Jampp
  • 内置的机器学习在线估计器每小时处理超过6000万条程序化广告信息(拍卖的中标率), the second price auction's costs, and new user and in-app events conversions).
  • 对模型进行优化和特征工程,转化率超过15%,公司净收入增加30%.
  • 开发收入和库存采购优化系统,使净收入增加20%以上.
  • 创建了多个web界面,为机器学习和优化系统提供可见性和可解释性.
  • 领导数据科学家定义和开发处理超过2亿用户的用户集群系统.
  • 开发了一个系统,使用前面提到的系统来做出各种决定——比如出价多少,为哪个客户——在实时拍卖中出售移动应用程序上的广告位.
Technologies: Pandas, Scikit-learn, Tornado, C, Presto, PostgreSQL, Bash, Git, Amazon Web Services (AWS), Linux, Python

Researcher | Developer

2013 - 2014
Integrative Neuroscience Lab
  • 使用Fisher的LDA和svm与EMOTIV EPOC脑电图耳机兼容,在Python中创建了一个思维拼写器.
  • Worked on a relaxation-based competition game using alpha brainwave detection among players.
  • 协助开发稳态视觉诱发电位选择界面,从远处控制乐高头脑风暴汽车.
  • Built a web server to control the headsets, launch the different systems, and record the brain activity and user input to enlarge our datasets.
  • 使用不同的分类算法和信号环境研究了念咒变体.
Technologies: Emotiv SDK, Python

Building Stronger Data Science Teams

Nowadays, working on a data science team involves working with people from different backgrounds, such as mathematicians, economists, actuaries, physicists, and engineers. 作为数据科学团队的技术领导者意味着你不仅要确保研究, insights, and products add value to the company but also that they are reproducible, maintainable, reliable, scalable, performant, testable, and correct.

For this reason, 我开始了一系列计算机科学和软件工程方面的现场技术培训,因为团队的数据科学家没有工程或计算机科学背景.

The training sessions were on topics like algorithmic complexity, programming exercises, environment setups, data structures, object-oriented philosophy, and software architecture.

On the tech side, 这与适当的入职和代码审查相结合,导致了更好的产品和更快的开发时间. On the social side, 我们很高兴地发现,通过学习和讨论共同的兴趣,我们意外地组建了一个更强大的团队.

Dynamic Spend Model for Digital Marketing

In the digital marketing industry, 绩效营销平台及其客户必须定义广告活动的目标和支出模式,以确定如何向客户收费(如固定的CPM), CPC, CPI or CPA).

我研究了一种新的消费模式,在这种模式下,客户和平台就最大预算和目标达成一致,比如每次点击或购买的最大成本,然后系统在实现目标并确保双方都满意的情况下,找到向客户收费的最优价格.

Event Predictions in Mobile Applications

我建立了一个统计学习系统来估计移动应用程序用户在应用程序上执行事件的概率. The online algorithm processes over 4 million data points per day to continually fit new data.

这是一个具有挑战性的项目,因为有不同类型的应用程序(如送餐或机票预订应用程序)。, different types of events (like searches or purchases), 延迟的转换反馈(人们通常不会在看完广告后马上预订机票).

Before deploying it to production and using it for revenue optimization, it was important to answer a number of questions:
• What is the minimum amount of data required to fit a good estimator?
• What happens if you stop learning from the data and keep using your current estimator?
• How much are you willing to wait for delayed feedback?
• Of all the events in an app which one do you want to predict?

After working on the problem and considering the possible contingencies, the system was rolled out successfully without any issues.

Financial Education Mobile Game

我设计并搭建了一款面向青少年进行金融教育的手机游戏后端.

The game consisted of a simulation of a rapper's career from the ground up, 在这款游戏中,玩家必须玩迷你游戏并做出正确的决定才能在职业生涯中取得进步.

The player's manager helps him by giving him advice and suggestions to balance leisure time, taking loans for playing in bigger venues and buying swag to increase the avatar's flow. The project was presented at a contest organized by Itau Bank and Red Hat.

Languages

Python 3, Python, SQL, Bash Script, Bash, C, HTML, JavaScript

Libraries/APIs

Scikit-learn, Pandas, REST APIs, NumPy, SciPy, Matplotlib, ZeroMQ, Keras

Storage

PostgreSQL, Redis, Amazon S3 (AWS S3), MySQL

Other

Click-through Rates (CTR), Machine Learning, Tornado, Algorithms, Distributed Systems, Optimization, Vowpal Wabbit, SSH, Multiprocessing, Dash, SaaS, Forecasting, Programming, Image Processing

Frameworks

Presto, Flask, Bootstrap

Tools

Git, Docker Compose, Vim Text Editor, Emotiv SDK, IPython Notebook, Plotly, Prophet ERP

Paradigms

Data Science

Platforms

Linux, Docker, Jupyter Notebook, Amazon Web Services (AWS), Amazon EC2

2015 - 2019

Master's Degree in Computer Science

University of Buenos Aires - Buenos Aires, Argentina

2009 - 2014

Bachelor's Degree in Computer Science

University of Buenos Aires - Buenos Aires, Argentina

Collaboration That Works

How to Work with Toptal

在数小时内,而不是数周或数月,我们的网络将为您直接匹配全球行业专家.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

在24小时内获得专业匹配人才的简短列表,以进行审查,面试和选择.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring