CurryAI: Artificial Intelligence to help Indian foodies stay healthy

A Computer Vision Aided Indian Food Nutrition Calculator


CurryAI: A Sophisticated AI-Enabled
Nutritional Calculator

Tuning Multi-class Classification for Large Number of Distinct Classes

In my previous paper on CurryAI, I described a computer vision aided Indian Food Nutrition Calculator
The crux of this work is using deep learning, and particularly transfer learning, in image recognition to distinguish Indian Food dishes from their photographs. This work can further be used to calculate calorie and nutritional content of a particular dish by using the USDA Food Central database that maps a dish to its components and nutritional content.
Our Solutions

We innovate, optimize and disrupt with our applied solutions

The founders of, a social entrepreneurship movement for free computer science education, have incubated - an AI Lab to provide solutions for real world problems that can be solved via the application of advanced techniques of Artificial Intelligence and Machine Learning.

01 Anomaly Detection

We have developed statistical models and data mining techniques to observe and detect data points that deviate from the normal behavior and characteristics of the underlying dataset. These approaches have been applied in detecting anomalous consumer behaviors, weather patterns, cricket scores and pricing models.

02 Image to Calories

Our most comprehensive work in ML application has been done on recognizing the contents of your meal (specifically Indian Meals and Food) from image to food’s nutritional content, including calories. This work is inspired by Google’s research paper: Im2Calories: towards an automated mobile vision food diary.

03 Crime Analytics

Using publicly available data of reported crimes in New York City, we researched data and identified patterns of crimes in different boroughs of NY. Developed an applied data sciences based solution to predict crime occurrence and nature of crime based on analysis of historical time-series data across New York to provide analytics with actionable insights to law enforcement agencies. This project was done at Data Sciences Program at Columbia University.

04 Speech Recognition through Natural Language Processing

Developed solutions for speech recognition for language learning instructions using Natural Language Processing technologies at MIT’s LaunchX Program.

Shubh Samtani

Shubh Samtani is a XII grade student at The International School Bangalore (TISB). He is the creator of Curry AI, a sophisticated machine learning based application that uses deep image recognition to assess the nutritional value of food by taking a picture of the food.

Shubh is the global winner of MARRS International Spelling Bee Contest which had participation from over 250,000 children. Shubh is among the top programmers globally on HackerRank platform, which has over 14 million active programmers. He has been one of the youngest to achieve a ranking below 10,000 at an age of 14 years globally on the platform.

Shubh volunteers as a teacher at 0Gravity, a global movement to create awareness, structured training programs and communities for computer education for kids 10-15 years old. He has been primarily responsible for porting all coursework of 0Gravity online for thousands of students, and designing and delivering the course – ‘Solving Analytical Problems Using Python’.

For more than last couple of years, 2020-2022, Shubh has conducted and published ground breaking research, under the guidance of Dr. Suresh Bhagavatula from IIM Bangalore, on India’s National Education Policy (NEP) and assessing the impact on the Right to Education (RTE) among 400 million children in India during the COVID-19 pandemic.

Shubh has completed the niche and highly selective global program for high school entrepreneurs offered by LaunchX (MIT Program) in 2021.

Shubh volunteers as a teacher in his social entrepreneurship venture, 0Gravity, a global movement to create awareness, structured training programs and communities for computer education for kids 10-15 years old. Shubh conducted groundbreaking research spread over 6 months in 2020, under the guidance of Dr. Suresh Bhagavatula from IIM Bangalore, to assess the impact to the Right of Education (RTE) among 400 million children in India during the COVID-19 pandemic. His research was published by IIMB Research Gate, and is referred by policy and decision makers in governments at State and National levels, heads of Corporate Social Responsibility (CSR) in different Corporates, and by the school administrations across the country to adapt education system from the learnings during COVID-19 and create sustainable changes in education policies in India. Shubh is vice captain of one of the four houses in his school, TISB. He has also played for school soccer and tennis team for many years.

Krish Samtani

Krish Samtani is an undergraduate freshman at the University of California, Berkeley, pursuing a double-major in Applied Mathematics and Statistics. He has a great passion towards solving applied problems in economics, statistics, and mathematics through the utilization of data science algorithms. Krish is a social entrepreneur, having founded two non-profit educational startups.

Krish is the principal founder of 0Gravity, a movement to impart free practical applied computer science training to children at corporate sites around the world. was co-founded by Krish Samtani in 2016, a Bengalurean who was 14 years old then. Since then, the initiative has continued to grow and is now live across multiple cities in India, the US and the UK. The movement has provided hands-on coding education to over 1,000 children, including more than 500 girl students through their workshops. Krish is a passionate advocate to create awareness for free applied computer science education for underprivileged children as a way to give them equal opportunities for future in a highly technical world. 0Gravity movement has been appreciated by governments across countries, featured on BBC, ABC News and covered by multiple leading publications around the world.

Krish is co-founder of Nilo Learning, a cognitive product-based firm for language learning instructions using Natural Language Processing technologies. Nilo Learning was judged as the start-up with most traction at LaunchX 2019, the world’s premier most summer program for high school entrepreneurs. 

He has built expertize in data sciences and its application to anomaly detection. Krish researched and submitted a thesis on statistical analysis comparing anomaly detection models for large datasets, and building applications for data visualization in Tableau.

Krish has been interviewed live on BBC, ABC News, and other media channels, including many magazines and newspapers. He is excellent in teamwork and forging partnerships with people of varied backgrounds. 70+ volunteers working on 0Gravity initiative, including many NGOs and professionals from IT industry.

Send A Request

Got A Project Or A Partnership In Mind?