I decided to start this project after many interactions with a wide variety of companies. I realised the main reason why most ML projects fail to deliver value, or even go into production, is the lack of a coherent strategy that closes the gap between the 3 keys elements: business needs, technical capabilities and human resources. To achieve this goal companies need a strategy.
Another reason for failure is that current consulting companies are normally lead by people without the required business or technical skills – attributes hard to have combined. As a result, they end up being a costly and misleading advisory service.
I truly believe that a technological transformation can be accomplished with honesty and knowledge, not with hollow claims. This is why I created the data science strategy plan.
I have been more than 12 years in digital marketing industries, and have help more than 240 companies to grow their business.
I speak to corporate, government and non-profit audiences around the world. I have spoken for over 20 clients in 5 countries. My speaking engagements range from keynote speeches to small-group workshops with senior executives, conducted in-person or virtually over videoconferencing platforms.
My keynotes focus on the importance of AI to improve productivity and build competitive organisations using emerging technologies in artificial intelligence. Through use cases, I stress the dos and don’t of this transformation process and avoid common pitfalls.
I will show how to start this journey and evolve to more advanced orchequestration where ML and Data is at the core of the business, and how ML is different from software engineering.
My current speaking topics include
– How AI will transform business through the use of new value creation mechanisms.
– How to becoming a data driven organization
– What it takes to build a successful data science team
– Deep Learning for business innovation and digital business transformation – leveraging the power of the latest algorithms in organizations.
In recent months, I have been specially focused on life in the post-pandemic world – the future of work, learning and play.
Convolutional LSTMs (meetup Londres)
Oeiras exponential economy
Congresso de Medicina Interna
Conferencias em Londres
I’ve worked with organisations across all industries who have spent millions of dollars on technology, but because of the inability to meet the challenges of becoming an AI centered organization, have returned to their traditional business practices with big losses.
My speeches are focused around the following topics:
- Recent breakthroughs in AI (reinforcement learning, NLP, image processing, structured and unstructured data, knowledge transfer) and their impact on business.
- Cutting through the hype and understanding the fundamentals components in deploying a successful data science project.
- How to become a data centric company, better serve your clients and enable the virtuous cycle of data+ML innovation.
“Technology should solve problems, not create them”.
Strategic Data Science
With hands-on experience, I help organizations design and implement Machine Learning projects and design a Data Centric Strategy at several stages of development, from designing DS pipelines to training existing DS teams.
An enterprise data science project is highly complex and requires deployment of an interdisciplinary team that involves assembling data engineers, developers, data scientists, subject matter experts and business stakeholders. I help your company build a strategy and a detailed roadmap preparation to deliver impactful DS projects.
Contrary to many consultants, we can provide a hands-on consultancy from years of experienced, from NLP to image segmentation and classification (see our code)..
Beyond technical implementation, I help integrate the long-term strategic goals and define a realistic road map with expected outcomes.
Several types of Data Science:
- The cosmetic data science
- Incremental data science
- transformational data science
- Foundational Data Science
Data Science projects have to be designed very at business unities.
I will introduce to concepts such as:
- The Data Science pipeline
- The Data Virtuous Cycle
- The Scalability dilemma
- Humans in the loop: where automation touches manual processes
- Why simulations can be more important than reality
- The bottlenecks of Machine Learning innovation
- Precision is not everything
As a consultant, I focus on the following topics:
- Identifying the business needs and design an effective roadmap to deploy a ML project
- Help navigate the technological requirements to deliver value and integrate the outcomes into the business.
- Prepare the organization for the implications of ML in terms of process requirements
- Identify opportunities to improve or disrupt business using innovations
- Introduction to new methods and improving existing data science teams.
- Verify the data quality, usability and consistency
- Introduction to advanced machine learning methods
Strategic Data Science
Below are some projects where we collaborate (see link more detail and a detailed example from Generali – link) – from CV
- AI in supply chain (TESCO)
- Insurance risk models – DL
- Cybersecurity (Santander)
- Fraud Detection
- Credit Risk
- Bayesian Optimization
- Ad optimization
- Digital Marketing optimization.
Retargeting advertising optimization
Mainad works on retargeting advertising. They use Google Cloud to serve thousands of ads per second and up to 100 million ads every day worldwide with a revenue of tens of millions USD.
They want to optimize the efficiency of the campaigns by targeting and pricing properly the right users in order to have maximum traffic in the most cost effective way.
We develop a solution based on LSTM neural networks to predict clicks and sales based on past behaviour of million of users. The solution was tested and deployed in the Google Cloud pipeline and is currently in production.
We were able to increase CTR by more than 20% and sales by 18% under the same budget.
Medical image processing
Contextvision is a major provider of image processing software throughout the world. They pioneered solutions for X-ray and Ultrasound image filtering and enhancement.
Contextvision wanted to develop a solution for an ultrasound probe to automatically segment, in real-time, arteries and veins in arms and neck captured in ultrasound video.
We develop a solution based on Deep Neural Networks using a convolutional LSTM (CNN+LSTM) to automatically identify arteries and veins in a very noisy video with a speed of up to 10 frames per second (without reliance on GPU). The code is available on http://github.com/farquasar1/CONVLSTM
We were able to detect arteries and veins with a consistent dice score of up to 80%. The algorithm was tested and deployed in a US client (Bard).
DirectLine group is a major insurance group that handles more than 3 million policies with an annual revenue of up to £2 billion.
DirectLine required to develop a new pricing algorithm based on the latest machine learning algorithms to better price and underwrite auto policies.
We develop a solution based on Lightgbm and Xgboost machines, including several techniques to handle high cardinality categorical data and a new method that does not rely on modeling risk as a Poisson process.
We delivered a set of algorithms that, for extreme losses, like body injury excess, increased the Gini index by more than 20% while keeping the same binarized Mean Average Error on equal exposure bins.
Tesco stock management
Tesco is the major distribution chain in UK with a turnover well above £32 billion with a complex supply chain and stock management system.
The client challenged us to develop an algorithm to estimate the products on shelves and respective uncertainty – due to losses, waste and errors in the supply chain.
We develop an algorithm based on Dynamical Bayesian Networks to estimate the stock levels at each point in time until the next MST (Modern Stock Take) takes place – every 6 months.
The model was able to detect if lack of sales was originated in no demand or no products on the shelf. An estimate of saving around £80 million was achieved through reduction of stock routines and sales availability improvement.