Bengaluru-based Mukunda Foods was founded in 2013 by college students Sudeep Sabat and Eshwar K Vikas as a Quick Service Restaurant (QSR) the same year to sell South-Indian food. While initially, the business was doing fine, challenges popped up when the founders tried scaling up the business and opening more outlets. Like most QSRs operating before the kitchen-automation era, Mukunda Foods couldn’t maintain consistency across its outlets, and that automatically affected its customer base. 

This was the founders’ first encounter with market reality in the Food and Beverages industry– scaling is tough, but maintaining consistency and controlling operational costs can be tougher. This was when Mukunda Foods pivoted from a QSR to a kitchen automation provider. Its first product, Dosamatic, was launched in 2016 and is a fully automatic dosa-making machine. 

Today, Mukunda Foods powers clients across the F&B industry–cloud kitchens, fine-dining restaurants, QSRs and hotels. It has automated more than 3000 kitchens across 22 countries, including that of food tech unicorn Rebel Foods, ITC, Ola Foods, Wow! Momo, Good Flipping Burger, Biggies Burger, Carnival, Samosa Party, and more. 

In a conversation with Analytics India Magazine, Rakesh Patil– Co-founder and CTO of Mukunda Foods, takes us behind the scenes of a kitchen automation business, diving deep to understand how artificial intelligence (AI) and machine learning (ML) come into play in smart kitchens and can help QSRs maintain consistency across outlets. 

“Maintaining food consistency is challenging for reasons like high attrition rate and SOPs not being followed. Automation could successfully bridge this gap where machines cook chef’s recipes with zero SOP deviation. This ensures the same taste across all the outlets, and the dependency on skilled staff also goes down. Getting automation also brings operational efficiency to businesses and also helps brands to sell faster,” Rakesh said. 

An MTech degree holder in Product Design and Manufacturing, Rakesh has earlier worked with Hindustan Aeronautics Ltd., where he developed engines for TEJAS Aircraft. Rakesh joined the Mukunda Foods team in 2015 and is responsible for developing the tech products from scratch and identifying new technologies for the future-ready commercial kitchen. Additionally, he also oversees the company’s research and development team and participates in strategy, standards, specification and processes planning.  

Edited excerpts from the conversation: 

Rakesh Patil: I have around 35 people in my team, including the design team, embedded electronic team, Android and IOT team and the COE team (new product development, image process, AI/ML, prototyping, etc.).

Rakesh Patil: So far, we have automated around 14-plus processes, and most of these solutions are cuisine-specific and problem-solving in nature. Besides our flagship product Dosamatic, we have Wokie, the automatic wok used for Chinese food, Indian and Thai gravies; Rico is the automatic rice, noodle and pasta-maker; and Eco-fryer is used for frying french fries, momos, samosas, and burger patties in less oil. 

Our latest addition, an E-pan, is designed to counter challenges faced by food businesses concerning blackening. It helps regenerate frozen, par-baked, fresh food items like flatbreads, kebabs and patties in their fresh and original taste. 

Kitchen automation machines by Mukunda Foods | Source: Mukunda Foods

Rakesh Patil: We are making machines smart by using AI and MI technology, which help our clients save time, money, and manpower and deliver consistent food to their customers. 

For instance, we have been using AI/ML with image processing in Wokie to make multiple cuisines from the same machine. The machine captures an image of each ingredient, also weighs and maps with the pre-loaded recipe and makes sure the operator doesn’t deviate SOP and also alerts the operator in case of deviation. The machine is also connected to POS and reads the instructions provided by the customer (less oil or spicier), and automatically customises the standard recipe.

Wokie by Mukunda Foods | Image Source: Mukunda Foods

Rakesh Patil: We have been using TensorFlow, Pytorch, AWS(IoT), Firebase, and Postman in the backend for our AI/Ml and Image Processing.

In the frontend, we have developed multiple Android applications which help customers create their recipes, remotely monitor machines, and request services using the same application. Also, we have been using HTML5 for web page development to display all the analytics to the customer at a different level and plan the operation and stock at each store accordingly. 

Rakesh Patil: One of our clients, Wow! Momo, was facing concerns regarding the under and over-frying of the items. They were looking for a solution to solve this problem and save operational costs.

Mukunda Foods worked on the project for almost a year and came up with the automatic Eco-Fryer. The fryer enables the operator to select the recipe to be cooked; once the food is put in the basket, the fryer automatically maintains the pre-set oil-frying temperature, and the basket automatically gets dipped. So, the food gets cooked at a set temperature and time. Also, features like double-dip, sleep mode, and deep-sleep mode helps save electricity and oil cost by 25 to 30 per cent.

In addition to this, the AI helps us understand order trends, peak hours, etc. This also helped in maintaining the temperature of the oil, leading to less wastage and expenses. We are further working on integrating image processing to Eco-Fryer, which will help auto-select the recipes.

Rakesh Patil: Automation is surely here to stay and gradually be a part of every commercial kitchen. It might have taken some time for the Indian market to acknowledge it, but the pandemic gave the industry enough reasons and time to evaluate it. Moreover, the shortage of skilled staff and labour due to the migration, and with hygiene and safety becoming a priority in the post-pandemic world, automation adaptation increased. This got validated when not just large food chains but small businesses owners also started automating their kitchens. 

Rakesh Patil: Embracing the change and adopting technology with time is a must for all kinds of F&B businesses. Pandemic just accelerated this evolution, and traditional food companies are no exception. The industry witnessed technological advancements at different levels of the business ecosystem– supply chain and logistics segment, packaging of the food, and digitisation of the menu cards. 

Talking specifically about us, traditional food companies have always been a major contributor to our business. For example, stand-alone outlets and small chains form about 50 per cent of our revenue. 

Rakesh Patil: While we have installed machines across 22 countries, we have never taken consistent efforts to penetrate into the international market. Our immediate plan is to aggressively expand to the UK and US markets.

We have recently ventured into KAAS (Kitchen-as-a-Service), enabling brands to expand their cloud kitchens to a new location with a fully automated operational kitchen with CaPex and OPex-lite models. Our first kitchen became functional around two months back, and the response has been great, so the plan is to look at the expansion of KAAS. 

Rakesh Patil: COVID-19 did hamper the growth of food businesses, but things are gradually getting back to normal. We can already see a lot of F&B players raising investments and scaling, which imposes a big demand for food-tech startups.

The Indian food-tech market is significantly different from western counterparts, especially when it comes to kitchen automation. The solutions come with robotic arms and a lot of aesthetic value; however, we build more functional machines. The Indian market may not prefer higher-priced machines that can’t ensure return-on-investment (ROI). A food business can expect ROI in a maximum of six months with any of our automation.

Webinar Speed up deep learning inference 13th May

Conference, in-person (Bangalore) MachineCon 2022 24th Jun

Conference, Virtual Deep Learning DevCon 2022 30th Jul

Conference, in-person (Bangalore) Cypher 2022 21-23rd Sep

Stay Connected with a larger ecosystem of data science and ML Professionals

Discover special offers, top stories, upcoming events, and more.

ICA finds independent components rather than uncorrelated component.

GIT is pre-trained using the BERT encoder and KERMIT objective on an unsupervised LM task.

The basic tenet that Gato followed was to train using the widest range of data possible, including modalities like images, text, button presses, joint torques and other actions based on the context.

IISc plans to bring the Indian pursuit in this field on par with the rest of the world, with a dedicated and focused effort.

AIIMS Jodhpur will also deliver mixed reality enabled remote healthcare services in the district of Sirohi to strengthen medical facilities delivered to underserved locations.

The new Gaudi2 and Greco processors are purpose-built for AI deep learning applications, implemented in 7-nanometer technology and manufactured on Habana’s high-efficiency architecture.

Protected Computing will allow users to remove personally identifiable information from Google Search results.

The summit will feature talks, workshops, paper presentations, exhibitions and hackathons.

Curriculum learning is also a type of machine learning that trains the model in such a way that humans get trained using their education system

Google informs that AlloyDB for PostgreSQL was built on the principle of disaggregation of compute and storage and designed to leverage disaggregation at every layer of the stack.

Stay up to date with our latest news, receive exclusive deals, and more.

© Analytics India Magazine Pvt Ltd 2022