Vijay Daultani

Experience

• 

  Accenture

  Sept 2010 - Aug 2011

  Associate Software Engineer

  • Excellent learning experience for me; gave me a perception of a tester how he thinks when he tries to find small and big test cases which might have skipped from the developers mind while he was developing the application. • Now while designing, and developing I keep in mind how to make the software resilient to attacks and also motivates me to prioritise tasks in my bug finding strategies.

• 

  Indian Institute of Technology, Delhi

  Jul 2012 - Jul 2014

  Teaching Assistant

  • Assisting faculty members with research and classroom activities and grading papers.• Advising and supervising undergraduate students for their assignments and projects. Showing them the direction so that they can focus their energy and time on learning wonderful and core concepts of the assignments.

• 

  NEC Corporation

  Oct 2014 - Feb 2017

  Researcher

  • Experimenting deep learning with convolutional neural network on NEC's SX architecture, inventing ways for optimal hardware utilisation for acceleration.• Invented a new technique for optimal instruction selection of a compiler, filed a patent and initiated a project for it.• Profiled, analysed bio metrics application, modified data structures and proposed algorithms to suite the underlying hardware architecture.• Understand different optimisation phases of the state of the art compilers like icc, gcc and llvm. Understand the problems and find their solutions which when provided to the compilers they can generate code more suitable for the hardware architecture. 部分表示

• 

  University of California, Berkeley

  Nov 2015 - Dec 2015

  Researcher

  • Researcher for collaborative project between NEC and UCB. The project was related to giving a optimal advice to the users whether to choose the conventional elevator system or stair case or escalator, by calculating and comparing the service time for each transportation medium respectively.• Initiated, designed and executed as a lead developer, an IOT project with swarm lab of UCB.• Integrating several technologies like python, Java and Android, a complete prototyping project was accomplished in one month. 部分表示

• 

  Rakuten

  Feb 2017 - Jul 2019

  Fraud Detection: • Created deep learning (FFNN, RNN) based fraud detection platform for predicting fraudulent credit card transactions for Rakuten Card. Discovered optimal transaction sequence representation for RNN. • Headed deep learning based hybrid platform utilising text and images features for predicting fraudulent recipes for Rakuten RecipeText Summarization • Invented new unsupervised algorithm for text summarization, filed a patent & published research paper • Utilized technology for summarizing financial reports, blogs and news articles to assist investment analysts to make quick decisions for Rakuten Investment Management • Utilized technology to experiment and explore automatic generation of meeting minutes for Rakuten MobileMachine Translation: • Improving neural machine translation model for Rakuten.com (global E-Commerce platform) for translating item pages for cross border trading in 4 languages (Japanese, English, Korean, Chinese)Search Query Prediction: • Invented new supervised method for automatic generation of user search queries and filed a patent • Automatically generated user search queries were used to make models to deliver better user experience and product recommendationsSentiment Analysis: • Led project to perform Recurrent Neural Network based sentiment analysis for the reviews across platforms of Rakuten.Attribute Extraction & Normalization: • Led project to extract item tags (attributes) and normalizing entities (values) using semi-supervised machine learning algorithms such as CRF and RNN for Rakuten IchibaProduct Recommendations: • Initiated and led project to extract user preferences from images for better product recommendationsAI Trainer• Conduct AI Trainings for Technology Division and Businesses. 部分表示

  • Head of NLP (Natural Language Processing) Team

    Jul 2018 - Jul 2019

  • Deep Learning Researcher

    Feb 2017 - Jun 2018
• 

  Amazon

  Aug 2019 - Sept 2021

  Sr. TPM Applied Machine Learning

  • Make suitable tradeoffs to optimize time-to-market and clearly communicating goals, roles, responsibilities and desired outcomes across cross-functional teams• Scope complex projects and deliver simple, elegant solutions to challenging problems with significant customer and revenue high-impact business problemsAutomated Advertisement:• Co-invented new machine learning (sequence to sequence) algorithms to generate search queries for automated advertisement in paid search channels• Interacted with a broad cross section of the Amazon organization to clarify ambiguous issues and negotiate effective technical solutions between development and business teams, to ensure on-time product delivery• Demonstrated communication and interpersonal skills including relationship building and collaboration with a cross-function team across 13 countriesJapanese Tokenizer:• Co-authored internal research paper on novel metrics and to measure impact of segmentation on downstream tasks of information retrieval. • Co-authored internal research paper on novel neural network-based lexicon-free Japanese text segmenter. Delivered integration of the segmenter in 6 business use cases 部分表示

• 

  Rakuten

  Oct 2021 - now

  Group Manager