The Silicon Review
Mosaic Data Science was founded in 2014 under the observation that there was a deep need for top-flight data scientists to work with organizations to integrate AI/ML into their decision-making processes. The Founders felt that there weren’t many mid-sized companies with the breadth of advanced analytics skills necessary to make this powerful technology useful to businesses. Mosaic Data Science’s parent company, Mosaic ATM Inc, had been building custom technical solutions to solve some of the aviation industry’s most challenging data problems for years, supporting customers such as the FAA, NASA, DoD, commercial airlines, and airports. According to Chris Brinton, CEO of Mosaic Data Science, the company’s data scientists were adept at pulling data from multiple sources, rapidly prototyping different algorithms, validating the most promising approaches, and building production-grade decision support systems. So, the team at Mosaic Data Science felt very confident in building actionable AI/ML at a time when these terms were early in their hype cycle, and bucketed the company’s skills and solutions under the Mosaic Data Science name.
Since then, the company has grown quickly and expanded its vertical offerings to a wide range of industries while building out horizontal solution areas around some of the most advanced ML techniques. Mosaic Data Science helps customers in sectors like energy, life sciences, retail, and logistics, design and deploy deep learning-based solutions around Computer Vision, Natural Language Processing, Contextual/Generative AI, Graph Analytics/Network Science, and Neural Network-powered IoT. The company also helps its clients combat the data science skill shortage and ascend the analytics maturity curve so they can use these powerful insights to make better decisions that benefit all stakeholders.
Meeting the ever-growing needs of the consumers and Mission of Mosaic Data Science
Mosaic Data Science makes complex artificial intelligence and machine learning solutions actionable, explainable, and usable to any organization. Data science should be available at scale to all firms, whether you are just starting to think about it or already have an established team. Customers like Mosaic Data Science’s practical approach to themes like digital transformation, generative AI, and supply chain optimization. The company boils more significant initiatives into bite-sized proofs of concept that deliver value in weeks or months, not years.
Drew Clancy, VP of Marketing & Sales states that “There are two recent project examples that jump to mind that really speak to Mosaic’s mission. We’ll keep the customers’ names anonymous due to non-disclosures. A Series A hospitality software startup had determined they wanted to use machine learning to predict when someone smokes in a hotel room via a device they developed, thus building their competitive moat around AI. The company had witnessed promising results with test data but saw large swings in predictive accuracy when extending to live hotel environments. Due to the size of the startup and the investor-imposed development timelines, the company hired Mosaic to quickly build a high-performing model that could predict the presence of smoke in different room conditions. After a kickoff phase, thorough exploratory data analysis, and ongoing collaboration with customer subject matter experts, it was determined that the use case was prime for a state-of-the-art deep learning technique known as time series classification. It is doubtful that the customer would have reached this conclusion alone or without the help of highly experienced data scientists at the competitive rates Mosaic charges. They wanted to use AI, but needed help, as most do, in fitting the models into their data, processes, and desired outputs.”
He furthermore added that, “In another project, Mosaic was tapped by one of the largest, diversified manufacturers operating in the market today to build an AI-enabled assistant that searched thousand-page operator manuals of complex industrial machinery to return relevant information that was either spoken or entered into a search bar. Large language models have garnered much attention in recent months, but Mosaic has been using these techniques and models to build powerful natural language processing tools that automate & improve decision-making processes for years. The manufacturing organization had an internal data science team but lacked the specific skillsets and real-world experience required to tune the proper NLP models to reach their desired outcome, so it hired Mosaic to build a custom solution that boosts the customer experience by allowing them to return relevant operator manual information in a matter of seconds. What I like about both projects is that each customer, despite being fundamentally different in structure, size, product, and value prop, desired to use AI and ML to change how they make decisions. Each customer was enlightened that there are experts like Mosaic out there whose core business is to contract out innovative services to help them reach their goals. True AI and ML solutions require a trained hand in customizing the tech to the customer environment; there is no single arrow in the analytics quiver to solve all use cases.”
Engagement models of Mosaic Data Science
Mosaic is committed to long-term, mutually beneficial client relationships, reflected in its 90% return business rate. Services & Consulting can be hard to define for some, so the team at Mosaic Data Science has created these engagement models to help with onboarding.
Capability Development: Are you looking to build an analytics team, take your team to the next level, or identify the most valuable AI use cases? Mosaic provides analytics assessments, customized training, evaluations, mentoring, and strategy consulting/road mapping to help companies find their footing in today’s fast-moving AI world.
Rent a Data Scientist™: Buried by projects? Going through a hiring freeze? Looking for a new analytical perspective? Mosaic’s trademarked program matches the proper onshore, full-time data scientist(s) with a client’s need.
Project-by-Project: Let us cross that project off your to-do list. Mosaic develops initial Proofs of Concepts through production-grade data science applications, and everything in-between.
Model Deployment & Scale: Deploying models into production with the necessary data pipelines often requires a different skillset and expertise than algorithm selection. Mosaic maintains data scientists and software developers that deploy models in production environments with MLOPs practices.
Insights by Mike Shumpert, Managing Director of Data Science for SMEs
There are quite a few low-hanging fruit opportunities with AI & ML. The open-source tools really foster a community, and there is a wealth of information out there to make AI/ML accessible. Automating and improving forecasting power by integrating more variables, testing different algorithms, and scripting can save small and medium-sized businesses time & money with improved visibility and better operational planning. OpenAI (ChatGPT creators) has helped the world start to connect the dots on what is possible with AI, training a deep learning model to automate a simple language or vision task is not as hard as it used to be. There are still large risks with not being able to understand what is happening with your model’s mechanics, especially as you think about extending a trained model to data it has not seen.
When The Silicon Review asked some examples of how Mosaic Data Science’s customers have found Mosaic Labs useful, Chris Provan, Chief Data Scientist stated that “Mosaic Labs represents our investment in hiring data scientists that specialize in deep learning. Many often think of Artificial Intelligence as the insights generated and recommended by using the automation benefits of Neural Nets, Computer Vision, NLP, etc. and the recommendations provided by layering optimization tools to prescribe an action for a human to take. A few years ago, we started to see a shift in the market as more and more prospective customers wanted to use these techniques to solve high-value use cases. “
He furthermore added that “Mosaic jumped on the opportunity and now has hundreds of success stories leveraging these powerful algorithms. We view them as horizontals, the custom solutions we build to solve use cases that can be extended to a different industry. Our ability to efficiently and effectively formulate a Deep Learning or Optimization solution based on real-world applications leads to streamlined development, helping our customers find answers in weeks and months, not years.”
Meet the leaders behind the success of Mosaic Data Science
CHRIS BRINTON CEO | Founder
More than two decades of R&D experience in logistics. Founded Mosaic Data Science in 2014 and Mosaic ATM in 2004. Leads numerous analytics projects focused on improving the efficiency and safety of the National Airspace system. MS Electrical Engineering, Stanford | BSE Mechanical and Aerospace Engineering, Princeton
MIKE SHUMPERT Managing Director Data Science
Over two decades of product/service development across diverse industries. Solid record of applying data science to produce double-digit revenue growth and cost savings in both small and large companies. Formerly VP of Analytics for Software AG and VP of Global Product Management for Dun & Bradstreet. MBA, Georgetown | BS Systems Engineering (Operations Research), University of Virginia.
DREW CLANCY VP Of Marketing & Sales
Marketing professional with 10+ years of experience in helping companies grow through innovative marketing program design. Sets marketing strategy and executes that plan through a test and learn methodology. Interacts with prospects and clients. MBA Marketing Management, Business Analytics Syracuse University | BA English/Journalism, University of New Hampshire.
CHRIS PROVAN Chief Data Scientist
Over 10 years successfully leading and executing advanced analytics projects for customers in a wide range of industries. Organizational analytics strategy coach and data science training instructor. Expert at integrating advanced statistical, optimization, and computer science techniques into decision processes to drive measurable impact on business outcomes. MS Operations Research, Cornell | BS Math Vanderbilt | INFORMS Certified Analytics Professional