Jijnasa Nayak

Jijnasa Nayak

San Francisco, California, United States
2K followers 500+ connections

Join to view profile

About

I am a Software Engineer working at Oracle Cloud Infrastructure (OCI). Currently, I am…

Activity

Join now to see all activity

Experience

  • Oracle Graphic

    Oracle

    Redwood City, California, United States

  • -

    Redwood City, California, United States

  • -

    Los Angeles County, California, United States

  • -

    Los Angeles County, California, United States

  • -

    Texas, United States

  • -

    California, United States

  • -

    Greater Los Angeles Area

  • -

    Taiwan

  • -

    Kharagpur Area, India

  • -

    New Delhi Area, India

Education

  • University of Southern California Graphic

    University of Southern California

    -

    Relevant Courses: Artificial Intelligence, Analysis of Algorithms, Natural Language Processing, Database Management System, Information Retrieval, Software Engineering

    1) Winner of TechStart Global US -2020, Local Hack Day:Share by developing a product to help in the time of COVID-19.
    2) Taught coding in the middle schools at Los Angeles by being associated with the Coding School.

  • -

    Relevant Courses: Data Structure, Operating System, Computer Networks, Data Mining, Computer Organization and Architecture

  • -

  • -

    Vice-captain of the school, participated in various co-curricular activities.

Licenses & Certifications

  • DO Your Venture: Entrepreneurship For Everyone Graphic

    DO Your Venture: Entrepreneurship For Everyone

    edX

    Issued

Volunteer Experience

  • The Coding School Graphic

    Teacher

    The Coding School

    - 6 months

    Education

    Taught coding classes in middle schools at Los Angeles by being associated with the Coding School.

Publications

  • Detecting Data Leakage from Hard Copy Documents

    IFIP International Conference on Digital Forensics

    1) Built a model in which different QR codes were embedded (by manipulating the font pixels) in the copies of the same documents such that these are imperceptible to human eyes.
    2) Retrieved the QR codes using the following methods with the corresponding accuracies OCR (84%), SVM (87%) and KNN(76%).

    Other authors
    See publication

  • Syntax and Rule Based Machine Translation from English to Odia

    International Conference on Machine Learning and Data Science (ICMLDS 2018)

    1) Developed a syntax-based and rule-based technique to generate the translated text by constructing rules based on the sentence structure in the input and target languages.
    2) Mapped the English words directly mapped to the Odia counterparts (obtained using web-scraping) and formed the remaining words (for which direct translations were not available) using Odia phonetics.

    Other authors

  • UHop: An Unrestricted-Hop Relation Extraction Framework for Knowledge-Based Question Answering

    North American Chapter of the Association for Computational Linguistics (NAACL-HLT)

    In relation extraction for knowledge-based question answering, searching from one entity to another entity via a single relation is called “one hop”. In related work, an exhaustive search from all one-hop relations, two-hop relations, and so on to the max-hop relations in the knowledge graph is necessary but expensive. Therefore, the number of hops is generally restricted to two or three. In this paper,
    we propose UHop, an unrestricted-hop framework which relaxes this restriction by use of a…

    In relation extraction for knowledge-based question answering, searching from one entity to another entity via a single relation is called “one hop”. In related work, an exhaustive search from all one-hop relations, two-hop relations, and so on to the max-hop relations in the knowledge graph is necessary but expensive. Therefore, the number of hops is generally restricted to two or three. In this paper,
    we propose UHop, an unrestricted-hop framework which relaxes this restriction by use of a transition-based search framework to replace the relation-chain-based search one. We conduct experiments on conventional 1- and 2-hop questions as well as lengthy questions, including datasets such as WebQSP, PathQuestion, and Grid World. Results show that the proposed framework enables the ability to
    halt, works well with state-of-the-art models, achieves competitive performance without exhaustive searches, and opens the performance gap for long relation paths.

    Other authors
    See publication

Courses

  • Analysis of Algorithms

    CSCI 570

  • Applied Natural Language Processing

    CSCI 544

  • Database Systems

    CSCI 585

  • Foundations of Artificial Intelligence

    CSCI 561

  • Information Retrieval and Web Search Engines

    CSCI 572

  • Operating Systems

    CSCI 402

  • Software Engineering

    CSCI 577A

Projects

  • CitizensHive: Matching Caregivers with Senior Citizens

    -

    1. Designed and developed an online platform to match senior citizens with caregivers based on location, fees, reviews etc.
    2. Implemented APIs for filtering, messaging, giving reviews, making payments (stripe API) and created user dashboard.

  • Skin Lesion Segmentation and Classification

    -

    1)Built a convolution-deconvolution network using dilated residual layers in the encoder side and pyramidal pooling layers in the decoder side for the skin lesion segmentation which achieved an accuracy of 93%.
    2) Devised a model for binary classification of skin lesions (melanoma detection) which used the pre-segmented images for better accuracy and built an m-way classification based on severity using appropriate feature selection methods, one-against-all SVM.

  • Neural Style Transfer

    -

    1) A CNN network was devised to separate and recombine the image content and style of the images. An algorithm to synthesize new images having the content of an arbitrary photograph with the style of a well-known artwork was implemented. This concept was extended to videos as well. The colour of the content image was preserved during the synthesis of the new image at a later stage.

  • Machine Translation from English to Odia.

    -

    1) Developed a syntax-based and rule-based technique to generate the translated text by constructing rules based on the sentence structure in the input and target languages.
    2) Mapped the English words directly mapped to the Odia counterparts (obtained using web-scraping) and formed the remaining words (for which direct translations were not available) using Odia phonetics.

Honors & Awards

  • Winner of Local Hack Day: Share Hackathon

    https://share.devpost.com/

  • Winner of TechStart Global US-2020

    https://techstars.org/

    Developed a system for touchless temperature monitoring system when COVID-19 was at its peak.

Languages

  • English

    Full professional proficiency

  • Hindi

    Full professional proficiency

  • Oriya

    Native or bilingual proficiency

More activity by Jijnasa

View Jijnasa’s full profile

  • See who you know in common
  • Get introduced
  • Contact Jijnasa directly
Join to view full profile

Other similar profiles

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Others named Jijnasa Nayak

3 others named Jijnasa Nayak are on LinkedIn

See others named Jijnasa Nayak

Add new skills with these courses

See all courses