102: Make the GRAIDE: Redefine the grading experience using AI with Graide

Holly Owens (00:03):

Hello everyone and welcome to another fantastic episode of Edup EdTech. My name is Holly Owens

Nadia Johnson (00:13):

And my name is Nadia Johnson. And we're your hosts

Holly Owens (00:16):

And we're super pumped today because we have a new and fun guest with us. We have man Gender, Kenneth, and he is the co-founder in c e o of Graide Man. Gender. Welcome to the show.

Manjinder Kainth (00:32):

Thank you for having me. We're

Holly Owens (00:34):

Excited to get into everything about Graide, and I've been doing a little bit of research as I told you before we popped on to the recording, so I'm excited to learn more about it. But before we jump into that, we definitely want to know more about your story. So how did you get here? Tell us about your journey getting into this EdTech co-founder, c e o space.

Manjinder Kainth (00:56):

Yeah, it's been quite an interesting journey. So I guess my journey starts by me being the son of two immigrant parents. My parents immigrated to the United Kingdom from India, and as a result, I haven't had a lot of financially accessed opportunities, but what I did get from them is this deep sense of grit and the idea that if you work really hard at something and you get a little bit of luck, there's very little you can't do. So as a result of that, I worked quite hard and did a physics degree and ended up doing a physics PhD looking at the mathematical and computational modeling of superconductors, and that was very interesting and we can talk about that if you'd like. But that was also the first time I had experience with assessment feedback. Now, I'd always done a bit of private tutoring on the slide.

(01:56):

I'd helped students do learn further exams when I was in secondary school and continued that when I went to university, being a crash course teacher for some students at schools. But this was the first time I kind of jumped into the assessment feedback space and kind of recognize the issues associated with that, specifically regarding the scale. You've got hundreds of students at university. So as a result of that, I kind of came up with a few ideas from a process point of view. My technical co-founder was doing a PhD in Vortex Dynamics in type two superconductors, got bored of that and decided to do, Whoa, that sounds complex a

Holly Owens (02:45):

Mouthful.

Manjinder Kainth (02:47):

It very much was. But then he decided to switch over to a PhD in education technology, and he thought about this idea of reducing workload and repetition associated with assessment, and we kind of combined those ideas together and built grade really. And then it's been quite an interesting journey since.

Holly Owens (03:07):

Oh, I love your story. I think it's great, and you've gone all the way with getting the PhD and everything, and I really admire that. That's not a easy feat or thing to accomplish for sure. So I love the story of your journey and glad that you're, and getting a little bit into grade, you're getting into the aspect of saving people time on the backend with grading and some of those workflows that instructors just have to go through. They're just so complex and we're alluding to here, it's just takes so much time.

Manjinder Kainth (03:47):

Exactly.

Holly Owens (03:48):

Yeah,

Nadia Johnson (03:50):

For sure. As a former educator, I can remember the ridiculous amount of time it took for me to grade. So giving that time back to instructors and educators is awesome. You told us a little bit more about your journey, but I'm interested to know who along the way has inspired you along this journey of coming into this space?

Manjinder Kainth (04:15):

Yeah, I think one of the things that's been quite interesting is hearing the stories of other entrepreneurs. So of course I mentioned at the start, I feel like I have this really strong sense of grit that was given to me by my parents, but from a operational and building a brand new thing from the ground up point of view, I've been quite fortunate to have people around me within the area that I live that have spun out small businesses and taught me the ins and outs of how to grow a business, hire teams, build products, and I guess most importantly, make sure that you're building something that actually helps people rather than just a project that you think might help people.

Holly Owens (05:07):

Oh, 100%. Yeah,

Nadia Johnson (05:11):

We

Holly Owens (05:11):

All have good ideas, right? Nadia,

Nadia Johnson (05:14):

Nadia And I come up with stuff all the time. We're like, we're going to do this, we're going to do this. But actually

Holly Owens (05:20):

In your own mind you think it's like, oh, this might be something that it's going to be a great solution, but then you find out, and that's the beauty of entrepreneurship. You can go test stuff out and then you can come back to the drawing board and say, okay, well this didn't work, so we're going to pivot in this direction.

Manjinder Kainth (05:35):

Exactly. I

Nadia Johnson (05:36):

Think I heard a quote the other day and it said, when you lose, you don't start from scratch, you start from experience. Oh,

Manjinder Kainth (05:46):

I love that. That sounds so good.

Nadia Johnson (05:48):

Yeah,

Holly Owens (05:49):

That is good. I'm writing that down.

Nadia Johnson (05:51):

Yeah, so I thought that was awesome. I was like, definitely you start from experience, you have an idea of where to go and once you start learning things, you can implement new things and just continue to grow from there. And I really loved that quote. That's just what this kind of reminded me of.

Holly Owens (06:10):

That's awesome. I really liked that quote. So being man gender, that you're in this ed tech space, you're a co-founder and a c e O, this is one of my favorite questions to ask on the show. We want to know how do you personally define educational technology?

Manjinder Kainth (06:30):

I wish I had a nice snappy five word answer to this, but I think unfortunately it's quite a nuanced thing because in some sense you could argue pen and paper is a form of education, technology, anything that facilitates the knowledge transfer from one group to another is really a technology and effectively pivotable. Now, we have done a lot of development in this space as a society moving from pen and paper to whiteboards and smartboards and tablets and laptops and software and all this fancy stuff. But the goal has always been, and in my opinion should always be making that experience of that knowledge transfer much more personal and much faster and much more effective, and that should really be the primary goal. So any sort of technology that actually does that is what I would personally call education technology.

Holly Owens (07:34):

I love it. So to the point and it makes so much sense. Yeah, I agree. One of the reasons I love this question is because we get a different answer every time, and they're all right, they're all correct, none of them are wrong. It's very open-ended and I love hearing it from especially people like yourself who are entrepreneur and c e O. It's great.

Nadia Johnson (07:59):

Yeah, for sure. I can agree with that, Holly. So now we're getting to the part where we kind of want to know a little bit more about Graide and how this service or product helps institutions and organizations within the education space. So if you could give our audience a little bit more information about Graide.

Manjinder Kainth (08:26):

Yes, of course. So Graide was built with the understanding that assessment feedback systems are pretty broken. For example, there are completely fragmented systems in place. So if you're doing a formative or summative assessment or you're doing an online or a paper-based assessment, you're using different systems. If you're creating your assignments in a particular place or you're delivering them in a different place or you're grading them in a different place, it just becomes a nightmare. And then there are, if you sit down and actually do the grading work, it's incredibly repetitive. It's you end up doing so much of the same sort of stuff. Now, it's not identical of course, because students have slight nuances to their approaches, but the approaches tend to be quite similar, and we tend to be grading those types of approaches. And of course the students are often dissatisfied with the experience as well. And when you take into account that workload is a leading cause of teacher exit, for example, in the uk, 40% of teachers quit in the first five years and the workload is the leading cause of exit. We have a similar statistics,

Nadia Johnson (09:50):

Ours is I think four years.

Holly Owens (09:55):

So we're right there with you.

Nadia Johnson (09:56):

Yep.

Manjinder Kainth (09:58):

It's a very serious problem. And I personally break down workload into two aspects. There's the administrative component and then there is the assessment component. Those are the two largest components when it comes to workload. And I think we're living in quite an interesting age because generative AI is here and I think it's going to make a huge impact on the administrative side. Sorry. And I think we're quite uniquely positioned to help with the assessment side. So hopefully we're living through a pretty nice transformational age with regards to that workload that I was mentioning. So what we decided to do at grade was break the problem down into two steps. The first step is to try to solve the more process-based problems. So let's make it so that the entire process is in an end-to-end platform. So if you're doing a formative or a summative or a paper-based or a digital assessment, it's all in one place.

(11:02):

It's the same place where you create the assessments, same place where you deliver them and the same place where the students attempt them and you grade them, et cetera. And all of this means that all those connection points between those different platforms doesn't need to exist anymore. So the only thing that really sits between the students submitting the assessment and them receiving the feedback is the amount of time it takes you to grade the work. We can do some really cool things like sum up all the points for you. I know that sounds like a really totally small thing, but if it takes 30 seconds to a minute for you to sum up those points and sync it into your learning management system, if you've got 30 students, that's half an hour. If you've got hundreds of students, we're literally saving hours right there. And that's just a small thing that just technology can really easily do.

(11:51):

And we wanted to make the collaboration much easier too, because consistency is such a huge issue when it comes to assessment feedback. So to address that point, you build a rubric on the fly and that acts as a single source of truth for you and your team. And what that means is that it's like having a Google Docs for your feedback. It means that everybody's always on the same page and you can always retroactively edit to make sure that moderation is much, much easier. And of course, you can do things like give extra students, sorry, to give students extra time or to be able to have multiple attempts and all these nice interesting things that you get from digital assessment. One of the other things that you get on top of that is extra analytics because you get interaction statistics. You can see when your students start the assignment, when they submitted it, how that interacts with the grades that they were receiving. They get much more detailed feedback that's directly relevant to their responses. You can type once and use multiple times. It's all designed to save you time from the process point of view.

(12:59):

Yeah,

Holly Owens (12:59):

I was going to say, when I grade, because I teach in higher ed, I teach grad master's level students and just grading papers and projects. I've tried so many different ways to figure out how to streamline, but I feel like at one point I solve one thing, but then it creates a bottleneck elsewhere. But it really sounds like grade just eliminates all that. And the saving time piece is so important to professors, educators because it allows for more time for us to create different learning experiences. And that's what I love about products like yours is that part of grading that's not necessarily like the fun part, but it's the required part that just becomes less of a worry, less of a concern.

Nadia Johnson (13:52):

And not to mention the students because the thing about feedback is if you don't give it in a timely manner, it's not helpful. The student needs to be able to get that feedback within a timely manner so that they can implement it within the next project or the next assessment or the next assignment. And so I think that also helps the student and it helps the teacher to get it done more efficiently, but also gives the student that feedback in a more timely manner in which they can actually implement the feedback and improve. So I think I really love this tool. It's awesome.

Manjinder Kainth (14:31):

Yeah, exactly. If we go back to Bloom's taxonomy, the timely component of that is really important. And if we also look at other key features for feedback, it needs to be relevant and it needs to be timely. Those two things we really try to address with grade now, we haven't even got to the cool part, which is the artificial. So this all kind of based on the idea that if something's repetitive, then there should be some space in some tech to try to reduce that repetition. You might have to be quite smart about how you do it, but there should be possible. Now we are in quite an interesting age of artificial intelligence and there's a lot of hype going around it. So I wanted to set the scene by saying, should we even actually expect AI to help in this sort of space?

(15:24):

There are two kind of big AI scenes right now. The first is of course, the generator models that we see in things like chat, G P T, and yes, maybe that'd be kind of cool. Maybe we can submit an assignment through chat G P T and see what it says. Maybe we give it a rubric. And if any of you listeners at home have had a chance to try that, you see that it does sort of work. But the problem there is it's not transparent, it's not auditable, it's not explainable, it is just a black box that does things for you and you can't fully consistently have that output. And that's not okay when we're talking about really important things like assessment, feedback.

Holly Owens (16:08):

Yeah, I think everybody's focused a lot on the chat, G P T and I use that every day.

Manjinder Kainth (16:17):

It's like a part of my morning routine with my coffee,

Holly Owens (16:20):

But there's other things that exist out there. Can you tell us a little bit more about how incorporates the AI stuff? I'm kind of interested to diving into that a bit.

Manjinder Kainth (16:29):

Oh yeah, absolutely. So the secondary type of ai, which has been arguably more popular but less impactful over the last few years is this concept of classification. So the way classification works is you have a giant bucket of data and it's labeled, you can imagine pictures of people and faces and objects, and you run that through a neural network, which is effectively like a fancy mathematical black box, and eventually you get something that can classify those particular objects. So does grading fit into that space? Well, not really, because grading doesn't have as much volume as those types of data. You often need tens of thousands of labeled data to be able to actually build that type of model to any sort of efficacy. And we don't really have that type of data. And the feedback that we give as educators tends to be quite personalized.

(17:34):

It's not just you don't want the bias to be propagated through different systems. You shouldn't be giving the same feedback to a high school student that you're giving to a university student, for example, even if you're giving them the same question. So naturally you might think that AI would have a difficult space here, but at grade we do something quite different. We understand that even though the variety and space of answers that students give can be incredibly large, the actual kind of true answer space in terms of the feedback is relatively limited. Students really only take a handful of approaches and paths. So provided that grade can take an answer and based on the feedback that you give, predict what that actual pathway is, it can kind of recursively learn in real time how to map between those two spaces. So very quickly in the space of three to 10 answers, the system can learn what type of pathways there are that the student is taking, and which ones are important to you as an educator to be able to suggest feedback based on your bank of feedback that you've already built up on how to grade other answers that it hasn't seen before, but look relatively similar.

(18:55):

And then because it suggests things, it's an assistant that's really important. It's an aid. That's why we spoke grade the way we did. Yeah,

Holly Owens (19:07):

I love it.

Manjinder Kainth (19:10):

It means that you can work with it to either accept or reject those suggestions and it will learn and update in real time. And that's basically how the AI works. What I find absolutely incredible about this process, however, is this concept that every single year we do a lot of grading work and all that effort is lost into the ether a little bit. Of course, the students get the feedback and the students get the grades, which is incredibly important, but it's just every single year it's very much the same thing. But this time, if you use grade two, go through your assessments, you actually are building a model that you can use for formative feedback in other years for other cohorts. You're actually investing the effort into building something that adds value into your process, and that's not really a paradigm that we've had before.

Holly Owens (20:01):

Oh, yes. Oh my gosh. I love this. And I want to go into so many different other questions, but I'd like to keep it open so that you can come back and tell us all the wonderful things that you're doing at Graide. And I, so I know this is probably really big in the uk, and I wonder if you have a, and you integrate with all the big LMSs that the big four, the Canvas, the Moodle, the Blackboard, the Bright Space by D two L. So I can imagine you have some interest over here in the States as well.

Manjinder Kainth (20:39):

Certainly. There aren't any pilots I can talk about publicly, unfortunately. Okay. That's

Holly Owens (20:46):

Okay.

Manjinder Kainth (20:49):

There are pilots running with relatively large establishments within the United States. Oh, I'm so excited. I hope mine's one of 'em,

Holly Owens (20:59):

But we will talk about that. Okay. So what we want to know now is you've done a lot, you're saving people time, you're giving students the feedback they need to be successful, this adaptability, this personalized learning situation. So is there anything that you can share that's on the roadmap for the upcoming year? What are some goals and things you have to get the audience excited about what you're doing at Greed?

Manjinder Kainth (21:27):

Absolutely. So one of the things that you might have looked at when you saw our website is we were predominantly focused on STEM assessments. So your engineering, your physical sciences, that sort of stuff, given the nature of those types of assessments. But I'm really happy to announce that we can now also grade essays and reports and much longer form content as well, which can of course make a huge impact. One of the really cool things that we can do is because a lot of establishments actually have this data in kind of a vault, we can take that data and use that to build pre-trained models that you can test out and you can use with your students to give formative impact right away.

Holly Owens (22:14):

Oh, fantastic. I love the fact that essays, oh, those take so long.

Manjinder Kainth (22:22):

Yeah, Exactly. And like you said, if you have 30 students, and let's say you're spending an hour, that's a whole day

Holly Owens (22:29):

More than a day. If you have a hundred students, that's going to take you a few days to get through that, but this really just streamlines it.

Manjinder Kainth (22:36):

Yeah, exactly. And then I'm also happy to announce that we're working on a proctoring solution, so if you want to run more high stakes exams virtually, you'll be able to run them at distance as well.

Holly Owens (22:54):

Yeah, good. I know, and I know a lot of people have a lot of feelings about AI and cheating and making sure students are turning in original work, so that will definitely help ease some of the anxiety and concern over that.

Manjinder Kainth (23:11):

Yeah, I mean, okay, I'll tell you one more thing. So we're also, yeah. Oh, Okay, good. Yay. So we're also working on something we're calling document integrity. So one of the really cool things about having the entire pipeline being controlled by kind of one system means that you have an interaction history with how the student interfaces with the document. So of course you can do simple things like see whether or not they copied and pasted something, but you can also do things like look at the cadence of typing and the edit history and see whether that interaction represents a genuine interaction. And it's a much more effective way to test whether or not that document that you've created is genuine rather than using an AI detector. Because unfortunately, AI detectors are very biased. They have huge false positive rates, and unfortunately they bias against international students quite heavily. So we're hoping to make a huge impact on that front because if you want to analyze assessments without generative ai, then you can use that, but then at the same time, you can allow your students to do it and you can see how they do use it and have a record of that and be able to give more interesting insight as a result of that. Well,

Holly Owens (24:33):

I'm so glad you shared that and I'm sure the audiences too. That's a good one. All right, so we're coming up here on the final two questions of the episode, and I've appreciated what you've shared so far. Last two questions we want to know, is there anything else you'd like to share? Did we miss anything that you want to share about grade? And then we want you to put your fortune teller hat on. We want you to look into the crystal ball and we want you to tell us what the future of EdTech looks like. So anything we miss and what does the future of EdTech look like?

Manjinder Kainth (25:07):

I don't think we really missed anything. I mean, of course we're always looking for more partners, so if you're interested, feel free to reach out and we can see if we can test things out.

Holly Owens (25:16):

Yes. So we'll put everything in the show notes so that if people are interested in partnering, they can click and go and find you.

Manjinder Kainth (25:23):

But in terms of the future of EdTech, I think we are living in a very exciting time. I think with the combination of generative AI and tools like ours, we can genuinely have an incredibly personalized learning experience. So I think in the not too distant future, we're going to have a system where a student can go through an incredibly personalized learning journey because if they want to learn a topic, we'll have knowledge graphs that define those topics. We'll have content that references those topics. We can generate personalized resources for them because we can have presentations, we could have videos, we could have audio all generated for us on the whim or podcasts for example, as well for students, if that's how they learn best, they could then consume that content. They could even have a conversation with that content using a chat bot. And then they could have bespoke questions that are tailored to them, asked to them, graded on a platform like ours where they get near instantaneous real-time feedback on their methodology, on their approaches and all the key formative important points to then attempt those assignments again, if required or if they've got the relevant proof points, get more personalized questions and move through that journey.

(26:42):

And then when they into a classroom and they interact with a lecturer or an educator, they can come in with a much more holistic point of view and ask those more awkward questions that the technology might not be able to interface with and still have that really high human contact without using Right now, I feel like sometimes we use teachers as a bit of information regurgitation machines, but that's not necessarily what education is. Education is a very human personal experience about creating those connections with those students. And this technology will really help facilitate that, and that's what I'm really excited about. Excited.

Holly Owens (27:21):

Absolutely. And we're excited too, and we would love to get updates from you again in six to 12 months to see what's happening and see the things. I'm really excited for the future of EdTech, and especially with companies like yours that exist, that are doing these awesome things in the space and helping educators and helping instructors and helping the learners. It's so important, and I can already tell that your company is very focused on what the end user needs.

Manjinder Kainth (27:52):

Yeah, that's really, really important to us.

Holly Owens (27:54):

Yeah. Well, thank you so much Mangen, for coming on the show. We really appreciate having you and sharing all the things about Graide. We'll have everything in the show notes where you can connect, where you can look up if you want to partner with Graide, and we appreciate all the things that you're doing in the support that you're giving people in the education space.

Manjinder Kainth (28:16):

Thank you for having me. It was a joy.