Natesh Babu Arunachalam – Senior Data Scientist at Finicity

The adoption of AI offers numerous potential benefits. However, it has also become increasingly common for AI models to pose their own unique set of risks. In this talk, I will present the various (potential) risks of AI along with how they can best be mitigated. I will also provide a practical pipeline for AI/ML model development, which intends to minimize some common risks.

Liliya Lavitas – Data Science Manager at Twitter

As one’s career is progressing they can consider transitioning into a Data Science Manager role. This transition can be challenging and overwhelming as People Management requires skills and training which many new managers are lacking. In my talk I will cover some tips and practical advice that helped me in my first six months of Data Science Management. This includes suggestions on how to set up expectations with your manager, stakeholders and with your team. How to prioritize your work and who to seek support from. All these recommendations are based on my personal experience as a new Data Science Manager. I will conclude the talk with a practical suggestion for those who are not sure if a Data Science Management career is the right track for them.

Adam Baron – Director, Big Data Quantitative Research, Data & Analytics, LSEG

StarMine leverages a diverse array of traditional financial time series and alternative data sets to create quantitative finance models for alpha and risk prediction. Quants essentially were data scientists before that terminology became fashionably. This talk with highlight a few research projects to show how data science can be applied to financial use cases leveraging graph network analysis, natural language processing, deep learning and alternative data.

Giacomo Vianello – Lead Data Scientist / Tech Lead at Cape Analytics

The development of AI and Machine Learning algorithms has sprung a host of new companies that are reshaping industries in many fields. For example, Cape Analytics is helping modernize the world of Insurance and Real Estate, by giving real-time access to unprecedented insights into the condition, risk and value of any property in the US and beyond. Applying AI to real-world problems at this scale demands state-of-the-art models, tools, and practices. It also presents technological and operational challenges beyond modeling that are often overlooked and can determine the success or failure of an AI effort. I will review these challenges and our solutions in the realms of computer vision and classical Machine Learning, from data collection and modeling to deployment.

Pragyansmita Nayak – Chief Data Scientist at Hitachi Vantara Federal

Data science projects have an astounding failure rate. Gartner 2019 states that through 2022, only 20% of analytic insights will deliver business outcomes. Technology advances are increasingly reducing the complexity associated with the development lifecycle and enabling convenient access to extensive amounts of computing and memory resources. Yet, the high failure rate is concerning. This talk will delve into five concepts that strengthen the foundation and will thereby increase the success rate – – Data Engineering – Data Ops – Data Catalog – Data Lineage – Multi-cloud hybrid Apart from the above five, another critical factor is the effective communication based on Design Thinking Principles between business and data team – aka, executives and data scientists. This will ensure that the right problems are being identified, prioritized and addressed in a time- bound manner.

Jeremy Hanna – Data Architect at Data Stax

Data, like physical mass, has momentum. The larger it gets, the more challenging it can be to move, redirect, and replicate. At the same time, many downstream consumers of data need timely access for preventing fraud, for real-time access, for personalizing and recommending, and so forth. This talk outlines strategies for managing this momentum among large scale distributed data stores, data in motion, and giving near real-time access to incoming data streams.

Linda Liu – Head of Data Analytics and Data Science at HyreCar Inc.

How can we systematically enhance the prospects of conducting conclusive A/B Testing to benefit the business? Everyone is familiar with the concept behind A/B Testing as it is simple and intuitive. Perhaps that is one big reason why A/B Testing is so wide used. It is one of the most important tools in data science and in the tech world in general as it is one of the most effective methods to draw conclusions about any hypothesis one may have. In some ways, it is a victim of its own popularity. It turns out not too many companies can carry out A/B Testing program successfully. Sure, everyone has a few winning A/B tests that he/she can be proud of and present at conferences. However, only select businesses have achieved real, long-term success via continuous and strategic experimentation. Why is that the case? The devil is always in the details. A holistic approach is needed to ensure external and internal factors are accounted for. A data scientist can play a key role in that respect. A data scientist’s role goes beyond merely measuring the results. I want to share how engaging a data scientist from inception to completion will systematically enhance A/B Testing. In other words, putting together an A/B Testing framework: from promoting an A/B Testing culture, to collaborating with various functions for idea generation, to identifying metrics to measure performance, to implementing the right technical approach, to monitoring, result intake, and possible follow-on operationalization, etc. I will also share examples on possible pitfalls and suggestions on how to avoid them so we can conduct conclusive AB Testing to benefit the business.

Dia Trambitas – Head of Product at John Snow Labs

Extracting data from documents stored as images, such as receipts, manifests, invoices, contracts, waivers, leases, forms, etc. is one of the essential operations many businesses still depend on. Up until now, extracting data from these images involved extracting the text through OCR and using NLP techniques, while neglecting the layout and style information which are often vital for document image understanding. Novel deep learning techniques combine features from computer vision and NLP into unified models, resulting in improved state-of-the-art accuracy for form understanding and visual information extraction. This talk shares real applications of these models to digitize and analyze documents with the purpose of extracting meaningful and easily exploitable data.

Mario De Felipe – Chief Data Officer at Grupo ASV

One of the main challenges of the company is to know how to get value from Artificial Intelligence and Machine Learning technologies. It is difficult for executives and owners of large and small companies to find use cases where they clearly perceive the benefit of using these technologies. In this seminar, we will address the main strategies and highlight the latest trends from analysts such as Gartner on studies of the so-called Infonomics.