PSL is making waves in machine learning and shows no sign of stopping! Most recently, PSL was listed #5 worldwide for machine learning services by Clutch.co. With over 832 organizations vying to be listed in the top 15, those that make the leaders matrix are among the most skilled in the discipline and most likely to successfully deliver on projects. Leaders are chosen for their ability to deliver on client expectations, which is determined using interviews conducted by Clutch with clients, market presence, and the caliber and strength of the company's clients. Clutch has crafted a set of evidence-based metrics that are regularly updated to ensure purchasers have access to the most accurate and timely information, allowing them to make informed decisions.
Over the years, PSL, and Colombia in general, have become home to a thriving machine learning industry. From the Colombian Ministry of IT and Communications project that will train over 4,000 students in machine learning, deep learning, and AI to the introduction of Holberton School of Engineering in Medellin, Cali, and Bogota and the opening of the Center for the Fourth Industrial Revolution and more, the machine learning environment in Colombia is well grounded.
But, as the fields of AI, machine learning and deep learning continue to grow in Colombia, so will the need for skilled talent to work and innovate in those areas. The good news is that while machine learning is still a relatively new field of study in Colombia, NGOs, local governments, universities and companies are quickly rising to the challenge of educating and boosting local talent. For example, PSL recently introduced a partnership with Holberton Medellin, a software engineering school where students complete a comprehensive, 2-year project-based curriculum and only pay for their education when they have secured a job. It's a new model of education that expands access to jobs in technology and boosts the local talent pool for companies.
PSL's experience in machine learning is already expansive and still growing. Even more importantly, PSL is consolidating and expanding the knowledge and expertise gained through various machine learning engagements in order to bring those practices company-wide. The process of institutionalizing machine learning is being addressed through training and development resources for employees, POCS in our R&D lab, Shadows placed on projects to encourage skill development in real world situations, and much more.
Recommendation Engine Driven by Big Data
Main components: Big data backend, machine learning algorithms
PSL is overseeing the development of a programmatic advertising recommendation engine dashboard guided by machine learning models. The purpose of the engine is to make real time bidding recommendations for advertising deals by using available data from the marketplace platform. The recommendation engine has been so successful that a lead data scientist from PSL has been training members of the client team on machine learning principles.
Algorithm Development for Customized Data Distribution
Main Components: Data lakes, machine learning algorithms, and data processing for big data.
PSL has been helping a client develop algorithms that use data freely available from the client to provide customized information to end-users. PSL created a process that first optimizes the data for storing, and then when it's needed, optimizes the correct data for use. By integrating spark and python machine learning algorithms, the client can use that data to predict special aspects for users. After extensive testing and training, the idea is to make the final models and algorithms available for purchase by 3rd party users.
Technology Stack: Kafka, Kafka extremes, data connectors, Spark, Batch Spark Streaming, building data lakes - HDFS, Apache Drill, Apache Phoenix, Magellan, Apache Zeppelin
Image Recognition for Passenger and Road Safety
Main Components: Deep learning, neural networks.
PSL created a real time alert, image recognition system allowing drivers to avoid external hazards posed by cars or other moving objects, while at the same time ensuring the passengers inside of the bus were safe by recognizing normal behavior. The team created a highly accurate neural network algorithm running on an embedded device with limited memory and footprint.
Machine Learning for Performance Engineering
Main Components: Performance testing, machine learning algorithms
The objective of the project is to automate the analysis of the data obtained from the performance tests carried out daily at PSL in order to improve the quality and accuracy of the resulting predictions. During the course of the project, PSL will define the machine learning algorithms that will be used to configure the results, including models for data preview and prediction that can serve as a key part of the strategy to allow us to better predict application behaviors in production and more accurately define the needed infrastructure for a particular level of demand.
With so many companies around the world looking to leverage machine learning solutions for different industries, it makes sense for them to start looking for nearshore software development partners with the expertise and experience needed. Colombia's mix of enthusiasm for AI, thoughtful and strong investment in the industry, and the increasing technical competency of local talent, create a very strong foundation for nearshore engagements focused on machine learning.
PSL in particular is continuing to emphasize and drive projects in the realm of machine learning, deep learning and AI, and our inclusion on the Clutch list of Top Machine Learning Developers worldwide is further recognition of those efforts and their success.
Ready to start realizing the benefits of machine learning? Let's talk.