Nester Joseph

Experience

• 

  VPhrase

  Jul 2017 - Jul 2022

  ⚙ Tools & Frameworks - • Python, Pandas & Numpy• PyTorch, hugginface, Flask, & ONNX Runtime• spacy, nltk & gensim• scikit-learn, XGBoost• OOPS & Data Structures• SQL, Spark• Git, Docker* A / B TestingI've worked on the following projects - ◆ Led the design and implementation of vPhrase's NLG System which is integrated with products like Phrazor (BI Platform) and Phrazor Plugin (for Tableau and PowerBI). ◆ Leveraging encoder-decoder language models (T5 and BERT) and word embeddings for building paraphrasing, text summarization and Named-Entity Recognition (NER) features in Phrazor and evaluating them based on BLEU and ROGUE evaluation metrics using Deep Learning Libraries (PyTorch, PyTorch Lightning, Huggingface).◆ Further optimization of trained models to increase inference speed and reduce the model size by leveraging concepts like knowledge distillation and parameter pruning using libraries like PyTorch, ONNX Runtime.◆ The design and implementation of Phrazor's Knowledge Base that is a huge relational database comprising logical relations between the KPIs, timeframes, entities, calculations and analyses used for insight generation relevant to various businesses.Key Responsibilities - • Finding novel ways of improving and scaling the incumbent NLG system.• Sprint Based Product Development.• Continuous Communication with the Product Management Team for new & existing features to be developed in in-house products.• Coordinating with the QA team for Bug Fixes and solving them promptly. Show less I've worked on the following projects - • The design and implementation of the first generation Advanced Analytics Component of Phrazor which involved predictive, diagnostic and prescriptive analyses along with outliers and anomaly detection.• Monthly MIS Report Automation use-case for a large Pharma Company based on Nielsen Data with Secondary and Tertiary Sales Performance Information resulting in 180+ hrs saved per month for internal analysts.⚙ Tools & Technologies - • Python, Pandas & Numpy• Scipy & scikit-learn• Excel• Tableau & PowerBI• SQL• OOPS Show less I've worked on the following projects - • Did various analyses of MIS data of industries like FMCG, Media, BFSI and Pharma Domain and compose them in concise reports.• The implementation of the report automation process of weekly reports on BARC data for a large Media company saving 9 hrs per analyst work of manually churning out MIS reports every week. • The implementation of a narrative driven dashboard based on internal MIS data for a large retail corporation appropriate stakeholders could log into and take actionable decisisons.• The analytics and implementation of an automated dashboard for performance reporting of an e-commerce FMCG company saving repetitive manual work done by in-house analysts. Show less

  • Data Scientist

    Mar 2020 - Jul 2022

  • Senior Data Analyst

    Apr 2019 - Feb 2020

  • Data Analyst

    Jul 2017 - Mar 2019
• 

  EXL

  Aug 2022 - now

  Data Scientist II