Akshit Sharma

• Currently tutoring Data Science students on various concepts of data science used in the industry.
• Making the learners industry ready by providing hands on experience to them on Python and various models used in Data Science.

• Developing a dashboard in excel sheet by automating the SQL queries using KPI Bees.
• Wrote a dynamic pricing algorithm to provide an edge in the market.
• Created a crawler for scraping the data of communities from Apartments.com to provide the data to the operations team.
• Currently working on fraud detection of coupons.

• Currently developing our automated AI payment Kiosk system for mass retailers.
• Responsible for developing the AI model used with Open CV and weight sensor for object detection.
• Responsible for the overall development of the product as well as raising funds for the company's operation

Worked closely with CTO on Following Products:

1. HT News:

Recommendation of news-based Timeline and Chronology of events (Potential IP with HT):

• Created a knowledge graph-based recommendation system for timeline and chronology. This has been filed for a patent with HT Media built on Neo4j.
• Lead a 5-member team and developed the data crawling infrastructure for HT news on an AWS Ec2 sequence.
• Integrated 22 news sites for crawling news and ingesting into the Created a machine learning model for Sentiment Analysis with 83% accuracy using Bert algorithm.

2. OTT Play:

Data Engineering:

• Lead the crawling team with Senior Engineering Manager and formulated a plan to crawl 50 OTT’s .
• Created a separate architecture to crawl 22 Singapore OTT’s and hosted it on an Ec2 sequence.
• Ran the downstream process on the data which consists of fetching the cast, crew and poster URLs for the movies and shows.
• Wrote an algorithm to re-fetch the missing IMDB ID’s, TMDB ID’s, corrupt posters and missing cast and crews for movie titles and show titles.
• Wrote a Django API for MSK server to store the analytics data from the front end. I implemented the producer MSK and consumer Kafka setup. The script was deployed on separate Dockers.
• Created and standardized a new crawling architecture to automatically run all the processes of a data crawler which involves crawling the data and running its downstream process. The new architecture is built on Scrapy.
• Created a de-duplication algorithm using Jaccard Probability to remove the duplicates records.
• Worked on Azure data bricks and wrote aggregation logic for various data jobs using spark-sql to run the recommendation pipeline.

Data Science:

• Created a movie recommendation model on Azure Data Bricks using Scala and Spark-Sql.

3. Slurrp:

Data Engineering:

• Wrote crawlers for crawling recipes and saving the data in the data base.
• Wrote an algorithm to download the recipes PDF’s from archive.org and save it in the S3 bucket.

Data Science:

• Created a model for classification of Vegetarian and Non-Vegetarian recipes using Bert algorithm: Used different features like calories, Nutrition content, ingredients in generating word embeddings from Bert and then feeding it into a classification algorithm.

4. MintGenie:

Data Engineering:

• Wrote crawlers for many different news sites to collect all the data regarding stocks of different companies.
• Wrote the deduplication logic using Jaccard similarity for classifying whether the News is same or different.
• Wrote a flask API and created the architecture to run the scripts on the AWS EC2 sequence.

Data analysis:

• Strategic Consulting and data-driven decision making for Business Growth & Optimization; Analytical Solutions to business problems.
• Solved business problems across product, sales, marketing through a data-driven approach.
• Worked on delivering meaningful insights to drive user growth, monetization, and sales channels optimization using the latest techniques.

Data Science:

• Developed full Scale and accurate Dynamic Pricing Models for automation of route prices.
• Developed full scale end to end data pipeline for dynamic pricing model.
• Created OBD Risk Profile

• Worked on NLP based project to analyze student feedback and automate responses on the VCP platform with the aim of providing better practice questions.
• Organized and managed 2 hackathons on regression and classification problems for 100 concurrent participants. The platform was designed on R Shiny.
• Mentored and helped a batch of 70 students in their capstone projects as well as placement preparation while also taking sessions on Python and Linear regression.