I develop efficient and scalable Machine Learning Models
As a Data Enthusiastic, I specialize in creating unique and captivating AI models and experiences for people and organizations who appreciate distinctive craftsmanship.
Hello, I'm Shubham Vishwakarma, based in Mumbai, India. With a deep passion for technology,
I'm on a quest to craft innovative solutions in the realms of AI and Data Science.
But that's not all; I've got a keen interest in the fascinating world of Software Development.
I'm excited to apply my skills, my drive, and my love for technology to real-world projects,
contributing to the growth and success of any endeavor I embark on.
Let's build remarkable solutions together!
Outside my regular developer routine, I enjoy playing cricket, watching Anime,
and listening to music 😉.
Investigated three months' worth of hotel data in Power BI and pinpointed causes for the market share decline.Estimated a 20% rebound in revenue and market share within a month
Python, Figma, PowerBI, Excel, Statistics
Designed a database comprising 17 tables, including authentication, account, and transaction types.Improved account administration efficiency by 55% to 60% for handling logs, transactions, and data.
T-SQL, Excel, Microsoft SQL Server
Developed a novel movie recommendation system that amplifies the experience by 60% to 70%. Utilized cosine similarity to propose movies based on user interests, reducing search times by 50% to 60%.
Python, Machine Learning, Statistics, Flask
Pioneered a time series forecasting model using Vodafone Idea Finance's five-year dataset, enhancing accuracy by 5% of existing models. Selected the SARIMAX model due to seasonal variations, resulting in a 4.5% MAPE inaccuracy.
Python, Statistics, Time Series Forecasting, Machine Learning
Engineered a voice-activated, emotion-based music recommendation system that reduced people's stress levels by 8%. Obtained an overall accuracy rate of 78% across all functions and duties of the system.
Python, Statistics, Deep Learning, Computer Vision, Flask
Implemented a LinkedList package for routine tasks, increasing productivity by 30% to 40%. Achieved a 40% decrease in code complexity by using basic LinkedList functions.
Python, Git
Employed an automatic encoder for image colorization using a down-sampled five-input layer and feature extraction. Escalate task accuracy and efficiency for colorization by 65% to 70% by employing a five-layer CNN.
Python, AI, Deep LearningFeel free to drop me an email if you believe we'd make a great team and would like to collaborate. Let's make our work together not only productive but also enjoyable!
Get in Touch