AgShift
Automated Food Inspection and Quality Assessment

According to the Food and Agriculture Organization of the United Nations, roughly one-third of the global food produce— approximately 1.3 billion tons—gets wasted every year. This food loss amounts to around $680 billion in industrialized countries and $310 billion in developing countries. The reason for this global food challenge is the traditional operational practices in the food industry. The industry today needs more transparency and consistency in assessing the quality of the food produced. But the lack of labor, innovative tools, and the prevalent subjectivity in food quality assessment poses significant challenges for businesses.

Founded in 2016, AgShift is on a mission to reduce global food waste by standardizing the production process through technological innovation. AgShift offers an autonomous food inspection system that is built on patented deep learning models and computer vision to inspect the produce for defects quickly and with high precision. In an interview with CIO Applications, Miku Jha, Founder and CEO of AgShift, shares her insights on their food inspection solutions and how they add value in the food supply chain ecosystem.

Can you give us a backdrop of the company’s inception and the passion that drives the company?

When California was hit by severe drought in 2016, many young and innovative companies were trying to address the irrigation problem and solve farming challenges using artificial intelligence (AI) and weather prediction technologies. During the same time, I—working as a senior executive at IBM—was leading many IoT initiatives. Having witnessed the drought challenges in the early part of my life, I was innately drawn toward solving this challenge. This drive led me to lay the cornerstone of AgShift. I realized that industries could provide more advantage to the stakeholders, including the growers and the consumers, by being more transparent and consistent in conducting the quality assessment on the food being produced. Today, AgShift makes quality assessment process better, effective, and faster by incorporating the best of machine learning and computer vision.

Tell us about the AI challenges you faced and how you overcame them.

The first challenge was to obtain a curated and labeled dataset of food images. To tackle this challenge, we tried to annotate the defects and change our data curation technique with the help of experts who knew exactly how to identify defects. Soon, our models started performing better.

The next problem was on the hardware side wherein we wanted to use a machine vision camera under controlled LED lighting; so, finding the right camera, setting the right configurations and integrating them with our software kit was painstaking. With our AI-based solutions coupled with the hardware setup, we could overcome these problems and improve our models and deliver accurate results quickly.
Explain in detail the working of the food inspection solutions that you provide to your clients.

Our AI integrated turnkey food inspection system, Hydra, identifies defects from the uploaded food sample images in an objective manner, and at scale. Hydra can be connected to any equipment within the facility. The system has a built-in software solution that is based on patented AI models. These models drive the inspection and identify the defects with high accuracy. Built on Google’s TensorFlow framework and library, Hydra helps in training AI to conduct defect detection on the food products automatically. This ensures consistency throughout the quality assessment lifecycle. At the same time, Hydra is extremely robust and works with the same precision even with large datasets. Eventually, it helps businesses gain more transparency and operational efficiency. Currently, Hydra is in the last phase of advanced commercial trials and can be expected to enter the market very soon. A lot of this vision has come together from our CTO, Dr. Amitava Bhaduri, who is an expert in embedded systems from Intel and leads AgShift’s global engineering efforts.

Besides, AgShift offers AgSkore, an app that allows users to perform quality assessment remotely using smartphones. Built to support organizations that are driven by manual processes, AgSkore allows inspectors to take images of the food product, enter them into the system, then leverage the system’s AI capabilities to analyze the data. With this app, users can generate food quality report from anywhere, anytime. AgShift’s solutions are enabling organizations to digitize and standardize the whole process.

We are very proud of our young engineering team of Vision and AI data scientists, who have put together two comprehensive solution - to disrupt decades old food quality assessment processes – making them objective, consistent and operationally faster for a better, more transparent tomorrow.

Give us insights into your client engagement strategy.

The key stakeholders in the food supply chain such as shippers, distributors, packers, wholesalers, and retailers can benefit from our technology solutions. With us, they can accelerate and improve the consistency of high volume food product inspection. Our team of experts with extensive experience across food and technology industries understands clients’ requirements and strives to resolve their issues cost-effectively. We ensure our clients have a faster, consistent, and objective food quality assessment. The company is also expanding its product areas by handling more food items in various verticals like fresh berries, edible nuts, seafood, etc.

What’s in store for AgShift in the future?

We are working on transforming every aspect of our business operations. On the solutions front, we are expanding our AI base to cover commodities like cashews and shrimps. Geographically, we are going very aggressive on the Asia market with trials underway with one of the largest commodity trading companies in the world for cashew inspections in their Vietnam plant. We are also scaling our R&D team and investing in harnessing AI talent rapidly.