Hi everyone. Thanks again for giving me this opportunity to talk about Rush program rush programming, how we can use this rush programming to build automation, enterprise level automation. Using this rush programming language, we'll also see as a technical exploration, we'll also see how we we can, how we can build a based AI solutions for transforming operational processes across the industries through high performance and memory safe implementation and architectures. Yeah. Hi everyone. Thanks again for giving me this opportunity to talk about rush program rush programming, how we can use this rush programming to build automation, enterprise level automation using this rush programming line. Which we'll also see as a technical exploration. We'll also see how we we can how we can build US based AI solutions for transforming operational processes across the industries through high performance and memory, safe implementation and architectures. As like memory safe programming language, and it will go through like how it is advantages for, and advantages and why it's very useful for, and or why it plays a very crucial role for building AI based applications. As we know rush is a programming language that is perfect for building this AI co-pilot because it gives us three main advantages. First one, it keeps our system safe from memory error. So that they don't crash or behave unpredictability, and it does not show it, it does not slow down during any processes or any high level throughput, right? It never slows down. That is the biggest advantage. And the second thing, it lets the huge amount of data to process at a time and without any problems and without any race conditions. And that means everything stays accurate and everything stays under a heavy load. That is the biggest advantage. And the third thing is it allows to build complex features without any extra performance cost. And in short, rust helps us to keep, make sure AI systems that are fast and safe and reliable. So when we look at the rust programming, right? As we know rust is becoming more and more popular these days and Rust is becoming a popular language or pop fab fabric language for many programmers as we know it's a memory safe and it's a low level code programming, right? And earlier days in future RAs might be place c C plus place or any other low level. System, system programming in my future, we, these days we see a lot of restorations, right? Let us see how how ra RAs can be utilized for building this copilot or AI based application, enterprise level applications, and how we can use the biggest features of rust and why it is advantages. We'll cover all of that. And that is a modern programming language that is becoming fabric for many many fab. Becoming favorite for many developers these days, right? Like the main advantage, the main reason for it is first one is it's a memory safe without memory safe. That means without slowing down the process it can handle many operations and, and also in many other languages, right? If you see memory errors and bugs that cause errors and that cause crashes, rash is designed to prevent all of that that is a biggest advantage. It's a memory safe. Any errors related to memory or any errors related to that are causing crashes, right? That can be Amen. That is the biggest advantage and, and it'll also it'll also prevent these errors without adding any runtime cost. So that is the biggest, that is one more advantage and coming to the second advantage, right? Why it is called fearless concurrency. The simple, it simply means like a lot of tasks can be handled at a time like multi multithreading. Basically multithreading when it comes to rest can handle this multithreading very simply and very easily. Massive sets, right? Like for example, multi-threading or multi-processing processes, right? Like ma handling, ma massive data sets or handling thousands of transactions or updating hundreds of systems at a time, or without creating data mix ups or conflicts. For example, if in a traditional programming to handle a thousands of transactions at a time, it'll be very complex for us. With the, if we are going the high level programming language like Java or C Sharp, it'll be very tedious job to handle. But when it comes to rest by default, that has that mechanism. At the root level, it is very easy for us to handle multi-level threading multi-level processes, especially this is helping this. Can help us in banking sector or financial domain because there will be mi millions of transactions per second, right? So Rust, that's why Rust is a fabricate language for building a blockchain. Technologies, which has lacks of transactions, a hundred thousand transactions per minute. And third advantage, right? Zero cost ab abstractions. We can build complex automation logic like decision making engines or real time analytics without slowing down without slowing down. The systems, for example, as right? Like when we are building the analytics systems, like BI reports, visualizations, like when it is processing when it is processing huge data loads to give the reports or do to show the analytics also show the graphs, right? We encounter a lot of vision. Those kind of challenges we can fix using this REST program language, which is one of the biggest advantage. So we put all together, right? Let us create a let us create this AI co-pilots using this rest programming language, how we can, let us see how we can create this AI co-pilots or AI safe application, right? AI applications using this stress problem, which can handle this high, but. Close and run without crashes. Deliver results instantly, which exactly what industry like finance, retail, and manufacturing and other utility utility companies are looking for. Let us look at it. The other slide. Let us look at the main building blocks that make the AI copilot work. First we have the data extraction engine. This is the system ICE and ERs. It collects all the information from different data sources like spreadsheets, documents, databases, or even for older file formats. It works fast and it can process ma multiple streams at once and does not waste the memory as we know. We can use this we can use so it can process ma, as I mentioned, right? It can process any. For any kind of files like it does not have any limits, right? It can process ESE file, audio file, or web file, or any kind file. It can be processing. And the next comes is the agent system. This, think, this is like a brain of Cooper. It uses the AI models to understand the data and make decisions, and it can also coordinate the task. It can even roll back actions if something goes wrong and it can add up. Based on the context. And basically this is a context, a system. I understand what exactly it is working on and what exactly the user is asking and what it can respond to. And even it can handle task automation, right? Anything really. It can perform some tasks. Like basically if there is a prompt that can not want execute any task, right? That can be handled by this agent system. This is like a brain that handles all the instructions. The main, core logic is present here. The final one is the integration layer. As we know, this is integration layer, which is used to connect to the different system like AI copilot if you want to integrate the AI copilot with the other integration systems, right? Even though it's a modern or legacy system, modern system that has API or or the legacy system, which does not have any a PS, it can work seamlessly and it can integrate with other systems safely and accurately. And these main, majorly, these three major parts of this copilot that that ma the major three parts, the data engine, AI brain, and the integration integration layer. They work together to make the whole automation system much more reliable and fast and easy to connect. Especially in the industry environment, especially where there are where there are thousands of transactions, right? For example, a financial industry or retail industry, which has thousands of transactions per second, especially. That is the, this is ideal in those words. We'll look at those examples in the next slide. We'll cover all of those use cases, real time use cases in upcoming slides. Yeah. One of the best things about, about using rust about using rust for copilot is that we don't have to keep keep reinventing the whole wheel again, once it is built built like a component, like a data parer or a reporting tool or addition making model, we can reuse it for a different project. It's like a module, like how we use the DLL files in C or regular. Regular programming language, right? We can use it like if we build one application, it can be used. Any component can be re imported or referenced into another project and reused, taken this is possible because the rust has very strong type system and it's a modular design. In simple terms, it means co the code is like a building blocks. It can fit together perfectly. Even if they are used in different business areas where we can create generic versions of tools, the, and then customize customize them, right? For example we are building, for example, I can give you, for example, we built a small. A POA system or a small application, that's for a finance application. It can be modularized or it can be it can be customized or it can be configured. For other industry. For example, we built it for finance and it can be configured for other industry also, or has, or healthcare or retail domain or other domain. Also, it can be modular as we know, right? ERP, like if we are building an ERP system like ERP system, like SAP using the rest, it can be modularized. If we have 10 features that we build for this ERP system and one company they want to use only five, and out of that five and other company want only three, it can be easily customized, as I mentioned. These are like, this is like a building blocks or building blocks the way the rust is built, the programming language is built, right? It's like a building blocks. It can fit all of them perfectly. That makes the rest rust programming very seamless and work and everything, all the components work together. And another benefit of another Ben biggest benefit is comp comply. Compile time guarantees. That means system checks system checks everything before even it runs. So if something does not match we watch it clearly instead of finding out later in production if even during the compile time only all the error are checked because in traditional programming or in the current programming world like usually we see a lot of production bugs, usually some of the issues, we encounter them only when we go to production. Those kind of issues can be analyzed and understood even before, during the compilation. We are using the REST program. This usability has a huge impact on time and cost in real world programs real world projects like companies have 60 have seen like 60 to 75% direction of development time for new automation workloads after they have reinvested once in f. Well, especially rest, right? As I mentioned, like you say, system system friendly language and low level prep. And there are many frameworks that uses this low code framework for building rest. There are on top of rest, right? They have built a subsets. Basically these are like subsets of basically the low code versions. We are planning to build any workflows, right? We can utilize these workflows for building new workflows like workflow engines or. Flows, especially business process automation. If you want to automate millions of transactions, or if you want doing any data pipeline or data orchestration jobs, right? We can leverage this RAs programming language, which is very flexible for building those kind of operations. So instead of speaking, instead of spending we spending weeks and building the weeks and months on spend building using these legacy applications legacy programming languages for development, right? We can adopt, to this even we have seen, as I mentioned earlier like we have seen a 62, 70 5% of reduction in development time. In build in development time for many companies, which started to use this rush programming and a fewer bugs, right? A major major advantage is it has a faster delivery, which is faster delivery, fewer bugs, main consistency, more. Consistent quality and quality across all consistent. If we are using us as a main Procore programming language for entire organization, right? As I mentioned earlier, we can reuse these components. One like if it is developed once, it can be reused anywhere. Let us look at how how this works in real time world. Especially with starting with financial financial services. One common one common example, is a loan processing and automation in the finance industry, right? The speed and accuracy is everything, and it makes it can make mistakes also, can lead to compliance issues or financial loss, especially you, we are using with the rush powered ai copilot can handle these thousands of transaction that time, right? Checking documents, verifying information, making sure everything meets the strict regulatory rules and all the validations, right? If we are going with this rest, this can be solved very easily because rest is a memory safe. It is ideal for handling sensitive data like personal details, income statements, or credit reports, or even we, or even without making any leaks or errors in practical world, right? Like we have seen document extraction has only 99.9% accurate, which means almost no mistakes. I have seen applications, right? Which uses rest for document extraction which, or OCR, right? Which has. Very, which are very accurate which that uses REST program, which are very accurate compared to the traditional traditional systems. The system also validates almost all the information even it's a real time. So the it's real time and so the approvals or rejections can happen very quickly. So if you go with the legacy applications or traditional applications it is, the process, like a process of doing the OCR process is handled by one system, and approvals is handled by another system. And basically the pre-process is handled by the one system and push process is handled by another system. But when it comes to arrest, everything can be handled together because. It can support any kind of application level. We can build any kind of applications using this. And moreover, we can integrate, for example we are building only the processing the core processing system in rest. And we want to integrate office 365 or core, our power platform, which has all. Built in approval systems, right? We can easily integrate this rust. So one part will be in rust. The major core part, the analysis part will be in rust, and the approval part will be in the other systems like power platform. And because of this because Rust is a very high performance system that usually runs smoothly even during the peak times, while hundreds of applications are in single hour, that means there is no delay to the customers and no backlog for the bank. And as we know rest it perform compared to if you see the benchmarks of a traditional programming, like a Java based application or a do net based application with rest, it's a low level program. It can perform very high high compared to high. It has high, very high benchmarks. It. In terms of performance that, that helps us especially during the high times, right? We have seen, especially in the banking sector sometimes the systems might go down because of the huge transaction. Especially, for example, if you take the trading trading world, sometimes systems might crash or systems might load. This can be used at this rest. Programming language can be used in such scenarios to handle very fast loads, especially without handling the crashes and handling the through handling those especially during the peak times. In short in finance, rush power, copilots can deliver perfect mix of accuracy, speed, and accuracy, and which exactly is needed by the current current finance sector. Let us look at the industry application. Second industry application like retail sector. How this copilot AI coate can be used for in retail industry, in retail in retail timing, and are critical because every small delays can lead to lead to empty shelves or mystery. Yeah, rush powered. The AI copilot can completely transform real time inventory management. It can handle distributed updates and meaning. That means it can keep track of stocks or across the hundreds of stores at the same time without risk of two system overriding each other. And. And this is possible because rush has a con, a concurrency feature that lets the processes of multiple stores multiple stores handle and update safely in a parallel parallel on top of that, right? Ai ai we can use ai. AI for analyzing the sales data and trends and parallel or, and make the predictive ordering or predictive suggestions. Basically, these are like supply chain for example. If you take for if you take in certain terms, in certain items, right? In certain items are selling faster than expected, the system may system can recommend of ordering of those items even before it, it ran out and, it does instantly instantly, no matter how large the data set is, it can process the data set and it can give us the predictor results. Another big advantage is legacy po OS integration. As we know already, most of the stores still use the legacy systems, legacy po o as a point of sale systems, right? Many retailers still use those systems. And more of, most of them does not have the re APIs, right? Like it cannot there. It's not they're not like iot, which can be integrated with internet or which can, does not have any internet connection or it does not have any APIs. We can have, we can build an integration layer that build, we can build these integration layer with those legacy systems using this rest, which can translate regular transactional data or p os data into extract the data into this hour. Based online system and do performance and do do a analysis on top of that as a result, right? As we know, we can work on legacy systems, it can be integrated with legacy systems. It is it can both ly do do, supply chain management and predict to analytics and real time operations, right? This makes RA based AI applications or RA based applications a best fit for retail world. Now let's talk about manufacturing and the utilities. The two sectors where down downtime and inefficiency can be extremely costly in manufacturing. A rest rest about ai. Copilot can automate I automate pro procurement and supply coordination. It can manage real. Orders or optimization. An example check a supplier. It can check the supplier availability. Compare the prices, placing orders in parallel without slowing down. It can perform even negotiations. Also even it can perform negotiations with multiple suppliers at the same time, it can ensure, quantity right quantity right quality or doc right quality. And also control the documents and automatically process those documents or process the communication, like whatever communication that happens during these processes, right? That can be handled by the same system itself. We may not rely on the other systems, especially, we have seen a situations where for supply chain management, they're using one system and for the other processes, sales and coordination, they're using outreach. They're using a different systems, right? Two different systems. Those can be avoided. If we are going with this rush based approach rush can replace those kind of, for example, in utilities, in let take the utilities world. In utilities many companies rely on world infrastructure like SCADA systems. Like Boulder Systems, flex Rush integration layer allows us to extend the system. These by, without replacing them, we can directly integrate them, as I mentioned in earlier, right? Like even license systems can be integrated with recession. That means without zero disruptions or zero operations zero disruptions to operations, like without having disturbing the existing operations, right? We can integrate, we can bring this in into our, into our ecosystem. We can we can add fault tolerance monitoring that keeps working even if. Part of the system fails as we, as I mentioned, right? This is a modular system. Even though some part of it fails, but the remaining part of it still can continue working. We can automate regulatory reporting for the compliance teams can, don't have to spend hours gathering the data. So by default as most of the modern enterprise applications have the reporting a log or transactional logs capability, right? The same way we can implement those kind of system, which can. Which can capture all the logs and all the transactions that are happening on these systems. Especially during the audit, when audit happens, right? If you want to if you want to see like why why what happened and all the digital stuff, the audit teams spending hundred percent hours of time on on understanding the data, analyzing the data, right? We can use this output as. We can save a lot of time for audit team. The final key benefit of using Rusty is it can handle large, complex operations predictably, whether it is monitoring thousands of meters in a power grid, or coordinating suppliers across countries, or the system remains fast and safe, reliable, and without any downtime. Now we have seen a couple of use cases, especially in utility world, finance world, and supply chain world, right? Now let let us look at the number numbers like how how rash is compared, like how can rash can perform compared to other, traditional approaches. First is the memory usage. Rushed ownership model means it uses about 30% less memory care compared to the regular or typical systems. That is important because it allows allows us to run the same workloads on smaller and more cost efficient infrastructure. Second one is a processing speed in real. In real world testing, RAs can perform data in incent intensive operations up to five times faster than compared to conventional platforms. This is a huge boost when it, when you are working on a high volume transactions or real world analytics. And the third third one is. Concurrency and efficiency that can handle a parallel task up to 200% more efficiently without risk for data corruption or race conditions. This especially especially this critical this is very critical in critical in industries where multiple systems are updating the same data at the same time. Especially in the finance world. These performance gains are, or just theoretical they translate into faster operations and lower, lower the cost, and also a smoother experience for both internal teams as well as the, and the customers. So beyond the, our performance Rest Restore also delivers a strong business and developer benefits. First one let us look at the first one. Computational efficiency with. Computational efficiency with zero zero overhead abstractions and a smart smart memory model. Rash can give us up to near near C level performance CC programming level performance while it is being safe in practice in practice. This means the 80% of 80% reduction in infrastructure requirements, fewer servers, fewer low fewer low cloud clause, and predictable predictable latency even when workloads changes. Plus also there is no garbage collection and pauses slow. And there is no garbage collection, which improves performance drastically. Next one is the developer productivity. Even even though Rust has a big learning curve, once team gets familiar with. Familiar with Thera programming, it saves a lot of time. We have seen 65% less debugging time because compiler ca comply. Comp compiler catches many error before, even before the other program runs. The carbo the cargo ecosystem manages to manages dependencies. The that makes e our developer life easy. And the documentation and the tooling are very excellent, as we know. It actually is evolving so much. These days. There are a lot of, a lot of resources for learning and it evolves so much. There, there are a lot of resources or a lot of tutorial available on web where once the team or the once the team gets familiar with this rush program it'll be. Very easy for us to implement this. Finally, infrastructure consolidation. Instead of maintaining the multiple specialized systems, a single rush code base can replace all of them as we know, right? Industries tend to use the different programming languages because of that, they have to maintain huge, different systems like for example, Java based system or Linux based system or windows based system. All of this, that, that can be this it can run on any kind of system. Infrastructure wise, we can eliminate those kind of limitations. That means unified monitoring, consistent development and consistent development patterns or the consistent and makes simple, makes the audit simple and fewer moving parts and fewer to manage and management wise, right? I it infrastructure, we, it is very hard to maintain different systems for different applications, which are not different ecosystems, right? So rash makes that if we replace those systems with rest it can, it can default those kind of infrastructure or maintena maintainability problem. So rust isn't fast. Rust isn't safe. Rust isn't safe. It also reduces operational cost, speeds up development and makes the system easier to maintain it. In long term. It saves lot of cost for the companies. Yeah. Now let's talk about how we can implement this rest for AI automation. Generally we, we follow a few proven design patterns that make our system both powerful and maintainable. For example actor model let's. Act for example, actor model that lets different parts of the system run independently and talk to each other without stepping on each other's data. But this is perfect for the complex workflows. And next one is the type state pattern, which helps us to lock the correct order or the step at the compile time, which so we cannot, we can't accidentally run the process in the wrong sequence. And the third one is command pattern. Which is useful useful for operations that might need that might need to underdone that might need to underdone. It pack packages each task so we can control we can roll back. Its, when it is needed. And another one is repository pattern repository pattern extracts how we can store and retrieve the data making it easy to switch the databases or backends without rewriting the logic. As we, as we use feature flags to enable or disable the features at the deployment time, which is great for testing the. Gradual rollouts. We can also leverage the powerful tools from rush ecosystem like Tokyo for handling. Lot of task at once and Saturday for working with different data coordinate REST belt, which is a natural language processing. And rayon for parallel data processing and tonic for G GRPC communication GRPC based communication. By combining all these data patterns the no known data patterns and the right tools, we can build automation systems very easily that are not. Only fast, but also safe and reliable, and which are flexible and adaptable. Let us wrap up let us wrap up the key takeaways from everything we have covered till now. First one is by investing in rest based automation framework, organizations can save up to 65 62 c, 75% reduction in development time for new workflows. And that is because once the core components are built, they can be reused across all the other projects and other departments. Second is memory, safety and pre and predictable performance that can lead to ni 89% of reduction in production incidents. That means fewer crashes, fewer bugs. And more firefighting for the development operation in terms. Third one is the overall process times overall process time can be improved by an average of 83%, which is directly translates into faster service delivery, faster services and faster deliveries, better customer experiences, and higher efficiency. So the message is simple, gives you. Gives you fast, safe, and reliable automation that can that can work across industries that can work across industries. It's a foundation that balances innovation and reliability, and it's a proven it's a proven to world, real world that the results in performance, cost savings, and quality. You for giving me this opportunity to talk about how we can leverage this s programming for building AI based to copilot. Now what next steps? Coming to next steps rather like the final conclusion? I want to say the first step is to evaluate your current automation challenges and see. Where the performance, reliability, or scalability are critical pain points, these areas are where we can fit the rest and show the biggest impact. Next is to consider the starting a pilot project. Choose the process that is important for your organization. Something where failure isn't an option. And test rush powered AI copilot platforms over there. And from there you can expand gradually reusing these components you have built in the pilot and to speed up the future projects. Once this is override, you can benefit ly faster development, like faster development, fewer issues, and more consistent results. The goal isn't to replace everything overnight, but it is about building the solid foundation where future proof of your automation given given you the PA performance, safety, and flexibility, you need to stay competitive. Okay.