Built on a belief that food is an amalgamation of art and science, technology is the main ingredient at Rebel Foods, a leader in India’s cloud kitchen restaurants. To begin with, the company uses machine learning to widen cuisine preferences, offers personalised food, and uses data analysis in areas such as supply chain management, recommendation systems, and more. By strongly integrating technology, it has built brands such as Behrouz Biryani, Faasos, Oven Story Pizza, and others.
Analytics India Magazine got in touch with Amit Gupta, Chief Technical Officer, Rebel Foods, to understand its strong tech hold and how it stands ahead of the competition using AI.
Built On A Strong Data Strategy
Gupta shares that technology & data are the backbones of solving industry-first problems at Rebel Foods. “Many decisions and light bulb moments for transitioning Rebel foods to a cloud kitchen model have happened on certain fact realisations out of data,” he said. It is in their DNA to make data-driven decisions and scale their offerings. “AI and ML are tightly integrated into every problem solving that are happening on the floor of our company,” he added.
The company is divided into three verticals — Food Discovery, Food Production, and Food delivery. Each of these platforms has its own problem statements, and most of these require some or other solutions on top of data & data science. They use software, robotics, and automation to solve food preparation at scale and provide consistent quality for diversified palates.
Some of the live projects at Rebel Foods are:
- Visual AI QC machines: that can detect SWAT (size, weight, appearance, and temperature) for each prepared dish and reject or accept based on the extent of deviation from ideal
- A robotics-led smart fryer: that gets adjusted automatically for the oil temp, dipping, and releasing based on what you are frying.
- Automated Wok: 100+ of their kitchens now have an automated Wok that dispenses oil, water, and other ingredients
Gupta said that they think of problems statements from consumers’ perspective first — “How can I be sure what I am eating? What hands have touched my food? What are the ingredients? I do not know of any restaurant that customises the food based on what I like or dislike? What am I allergic to? What are my specific dietary requirements?”
They further act upon these problems by using technology and data. “As we own the full stack control of the whole platform from SCM to discovery to order to preparation to delivery, we are envisioning the platform to provide complete transparency and control to customers for food order using any interface,” said Gupta.
When Data Science Meets Food: Use Cases
Data science plays a vital role in Rebel Food. They work on many exciting data engineering use cases such as personalisation, recommendation engine, dynamic pricing, predictive modelling for inventory, customer engagement, operating metrics for kitchen machines, vision computing for quality checks, sentiment and semantic analysis of user feedback, and more. “We are also launching face mask detection during order deliveries to keep a check on safety and compliance protocols,” added Gupta.
Explaining a few use cases around how new-age tech is used at Rebel Foods, he shared a few instances as below:
Download our Mobile App
Machine learning offers personalised food services: To bring about personalisation, they ingest large amounts of real-time and batch data and run data science models to improve customer experience on the ordering platform. They also use it to make the whole experience seamless. Machine learning is also used to make the discovery and ordering experience unique, fast, and seamless for all different types of users. As Gupta shares, “we want to go further to give as much control in the user’s hand as possible for food customisation and preparation by integrating the flows to backend systems (IoT).”
Data analysis for supply chain management: Gupta shares that they use tools such as Sap/Hana in combination with the in-house platform (Spark) for all the inventory and assets management. These systems are integrated with different vendors/partners to provide an end to end view of the whole process and workflows. He further added that they are using data science heavily in demand forecasting and inventory planning to minimise variance and wastage. Their models are trained on over five years of data to ensure the best possible accuracy levels.
Analytics-driven recommendation engine: As a tech-driven B2C commerce platform, the recommendation engine is one of the key modules which is owned and constantly improved by their data engineers and data scientists. They use various data types, on top of which they use a mix of content-based & collaborative filtering for recommendations. “Majorly, we measure the efficacy of our recommendation engine on 2 metrics – benchmark improvement of cross-sell/up-sell and relevancy of recommendation result-set,” he said.
Tech Stack At Rebel Foods
Gupta shares that they have built complete full-stack systems deployed on the cloud in food discovery, preparation, and delivery. They are a big believer in using open-source technologies wherever possible, and most of their technologies are built in-house. He further added that they have built their data pipeline using Kafka, spark, S3/Hive, and other apache open-source technologies. They also use data science libraries on AWS sage maker and some standard & some in-house built data visualisation tools.
The data science team also uses tools and algorithms in machine learning (regression, classification, random forest, k nearest, etc.), deep learning (Neural Networks, Tensorflow/Keras, etc.), AI, Statistics, SQL/NoSQL DBs, Semantic & Sentiment analysis, Predictive Modelling.
The company takes pride in its strong integration of technology, showing sincere commitment to redefining the way we eat and paving the path for the future of food while retaining the best of traditional culinary practices. “Though the brands and food choices could differ across the globe, the problem statements for the platform remain somewhat common, and we build solutions and platforms in that context, which makes us stand out” said Gupta in the concluding remarks.
What's Your Reaction?
Srishti currently works as Associate Editor at Analytics India Magazine. When not covering the analytics news, editing and writing articles, she could be found reading or capturing thoughts into pictures.