Tushar Abhishek

AI Researcher

Hyderabad, Telangana, India8 yrs 6 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in generative AI and NLP.
Published research at international conferences.
Experience in enhancing user experiences with AI.

Stackforce AI infers this person is a Data Scientist specializing in generative AI and natural language processing.

Contact

Skills

Core Skills

MlopsNatural Language Processing (nlp)Machine LearningSoftware DevelopmentMobile Development

Other Skills

Adobe FlexAlgorithmsAndroid DevelopmentCC (Programming Language)C++Cross-lingual fact-to-text alignmentData StructuresDataset creationDebuggingDeep LearningDynamic call routingFlaskGNU/LinuxGenerative AI

About

Hello! I'm Tushar, a Data & Applied Scientist II at Microsoft, where my primary focus revolves around enhancing the next-word suggestions in Bing Copilot Autosuggest using generative models. I completed my Master's degree by research from IIIT-Hyderabad in 2022, where my academic journey delved into the fascinating realms of text coherence, cross-lingual fact-to-text alignment, and text generation. During my time at IIIT-Hyderabad, I explored various aspects of text generation, particularly in low-resource languages and multi-lingual contexts derived from Knowledge Bases. My contributions have been acknowledged with publications at renowned international events such as WWW, PAKDD, ICON, reflecting my commitment to advancing the field. Transitioning from academia to industry, my current role at Microsoft involves being a part of the dynamic team working on Bing Autosuggest. My responsibilities include end-to-end model training, deployments, MLOps, model optimization, and staying abreast of the latest advancements in the field. I'm deeply passionate about the transformative role that generative AI plays in shaping the future of creative content generation and personalizing user experiences. If you share a curiosity for the exciting intersection of generative AI, natural language processing, and user experiences, I'd love to connect with you on LinkedIn. Let's engage in conversations and explore potential collaborations in this ever-evolving field.

Experience

Microsoft

2 roles

Data & Applied Scientist II

Dec 2023 – Present · 2 yrs 3 mos · Hyderabad, Telangana, India

MLOpsLarge Language Models (LLM)Deep LearningModel optimizationGenerative AINatural Language Processing (NLP)

Data & Applied Scientist

Aug 2022 – Nov 2023 · 1 yr 3 mos · Hyderabad, Telangana, India

Iiit hyderabad

2 roles

Teaching Assistant

Jan 2022 – May 2022 · 4 mos · Hyderabad, Telangana, India

Teaching Assistant for the Topics in Deep Learning course in Spring'22.
Course advisors: Prof. Charu Sharma & Prof. Makarand Tapaswi.

Research Assistant

Aug 2019 – Jul 2022 · 2 yrs 11 mos · Hyderabad, Telangana, India

worked on detecting the text coherence in long documents by fusing document-level information with factual information to improve coherence modelling.
worked on IndicWiki project for enriching Wikipedia content in Indian Languages using Knowledge Graphs (KGs) and Text generation models.
worked on the creation of dataset using Unsupervised cross-lingual alignment of Knowledge graph triple with a sentence in 12 Indian languages. Additionally used this dataset to fine-tune multi-lingual text generation models.

Text coherenceCross-lingual fact-to-text alignmentText generationKnowledge GraphsDataset creationNatural Language Processing (NLP)+1

Plivo

Software Development Engineer

Jul 2018 – Jul 2019 · 1 yr · Bengaluru Area, India

Improved dynamic call routing algorithm and developed pipeline for distributing voice call load uniformly across multiple voice-servers deployed in different regions.
Developed feature to support the processing of non-Latin scripts in voice call APIs.
Rewrote legacy Flask code in GoLang and refactored it to make the voice call APIs more extensible

Dynamic call routingFlaskGoLangSoftware Development

Ribbon communications

Software Engineer

Jul 2017 – Jul 2018 · 1 yr · Bengaluru, Karnataka, India

Developed module to monitor and debug the internal control block structure for software edition (swe) of Session Border Controllers (SBCs).
Worked on creating lighter version of swe SBCs for efficient IP packet processing using Intel’s Data Plane Development Kit (DPDK) library.

MonitoringDebuggingIntel DPDKSoftware Development

Syon research labs

Intern

Sep 2015 – Jul 2016 · 10 mos

worked on Fantasy Sports Guide Android and iOS apps using Adobe Flex. It’s an optimal lineup suggestion App for Football Sports.
involved in developing the Deepockets, a financial project (Android, iOS and Web App) that advises its users on utilizing their money efficiently.

Adobe FlexAndroid DevelopmentiOS DevelopmentMobile Development