roles

Discussed Skill Score Algorithm design changes and provided feedback and suggestions. Reviewed Back-End code for accuracy and correction.

Used Natural Language Processing (NLP) and Statistical concepts to develop the new Skill Score Algorithm for Talentsky v3.0.7. Implemented the new Skill Score Algorithm in Java for the backend. Wrote Python Scripts to test out the algorithm by ranking users based on their attributes.

Developed and ran a Python script to delete unsatisfactory images from existing production AWS bucket after having the production images reviewed by the image curation team.

Adopt Agile practices for efficiency, predictability and customer-driven product development for release 3.0.0

Run Python script written by me to fetch necessary images using keyword search through Adobe Stock API every month. Extract keywords from existing text data sources and identify keywords lacking images in our database and use them as search input for the script.

Closely work with technical manager and QA team to keep track of bugs and issues on JIRA board. Get a clear understanding of the issue and expected functionality and debug the errors and ensure the closure of tickets.

Worked on back end APIs in Java which involved fetching data by firing elasticsearch queries and processing them to return appropriate data response in JSON format to front end to support various features for flagship version 3.0.0 release of the product

Used state of the art embedding models and vector similarity methods to suggest Skill Mentor pages to users based on their profile information. Prototyped the functionality in Python and implemented it as a backend API in Java using Deeplearning4j.

Wrote a backend API which fetches the OG tags information for an input URL and returns necessary information like title, description, image URL to the front end to display to user resulting in better UI/UX. Implemented the API in Java using jsoup for web scraping and Caffeine cache to increase lookup speed.

Analyzed the task text data in Talentsky database to find tasks and topics having low count of images for suggestions. Used the Adobe images API to scan their image database to find images with tags giving a high cosine similarity score for the low performing tasks and topics. Wrote scripts to automate the image search and download processes and populated Talentsky's database thereby, improving the quality of image suggestions.