Sanket Shahane is Data Scientist at Lucidworks Inc. Passionate about machine learning, search, and the potential it has to transform businesses. At the intersection of business, machine learning, and field experience Sanket helps shape the Fusion AI product at Lucidworks. His focus of work involves researching and developing methodologies to solve complex problems of the search domain like the Cold Start, Developing Question Answering system at Lucidworks, and working closely with SOTA technologies in Deep Learning and NLP.
Sanket Shahane is speaking at the following session/s
Productionizing Python ML Models using Fusion 5
Lucidworks' Senior Data Engineer- Andy Liu and Data Scientist- Sanket Shahane will present on: How Fusion solves a major challenge Data Scientists, Machine Learning practitioners & enthusiasts face "Productionizing ML models". Today, Fusion has the capability to serve python based ML models at runtime without restricting the developers to a particular framework. Models developed using any python toolkit (TF/PyTorch/Sklearn) can be productionized within minutes. In this talk Andy and Sanket will present Fusion's capability to seamlessly integrate with Jupyter Notebooks, and show some of the advanced models being developed at Lucidworks; like Explainable Deep Learning classifiers (document categorization/content moderation), Named Entity Recognizers trained on custom datasets (text analytics), Reranking search results with Google Sentence Encoders, etc.
The audience will also experience the process of publishing an ML model to use in Fusion's query & index pipelines through a live demo.
Fusion's ability to 1. serve python ML models, 2. integrate with Jupyter Notebooks for accessing data from Solr as Spark/Pandas dataframe. 3. Process to publish and use custom trained ML model in Fusion's Index & Query pipelines. 4. Ways Deep Learning can power smarter search and data analytics.
ML practitioners looking to transform the search and data analytics space with the help of Machine Learning will benefit from this presentation. No prerequisite knowledge is necessary. Having basic idea about Fusion and some basic ML concepts will be helpful.