Yash Kolhatkar

So, uh, basically, I belong to, and I did my engineering from itself in electronics and telecommunication. Uh, after that, I did a post graduation diploma in big data analytics from. Um, after that, I started working as software developer in a company called SMS Magic. So SMS Magic company has basically has a a CRM product, and, Uh, there my main responsibilities were, uh, the like, uh, developing APIs, Integrating those APIs with a third party APIs such as Salesforce, Talkdesk. After that, uh, the main technology stack was, Uh, Python, grade, SQL, uh, this one. Uh, after that, uh, my my responsibilities include Uh, like, uh, production support and deployment as well. After, uh, I did some migration of the modules from Python 2 to Python 3. After that, I started working, uh, in a processing system as software senior software developer. Then I started working on the project called, and it was basically, Uh, serverless product. So we were using AWS for it, such as AWS Lambda for the, uh, API development, uh, as soon as for the, Communication after the and currently now currently, I am working in the, uh, Gen AI. So I am Integrating and leveraging capabilities to, uh, automate and exploit the process of the, uh

Can you explain the role of asset properties, uh, in PostgreSQL or any other database transaction and how they did aid in the database management. Certainly, ACID stands for Uh, automaticity, consistency, isolation, and durability, which are the set of properties that ensure level processing of database transaction. Admittedly, this property ensures that the transaction is treated as single in, uh, uh, indivisible unit of work. So either all the changes made by the transaction committed to the database or none object. Consistency transact uh, inconsistency transactions being the database from the 1, uh, one transition state to another. Database must satisfy, Uh, a state of integrity constraint, we can save. Isolation, this property ensures that multiple transaction can concurrently without interfering each other. So durability, once this transaction is committed, uh, once the transaction is committed, These changes are permanent and survive any subsequent failure. So these are the database management systems.

What are the subfiling leverage commonly used in finance, and how can they how can they be utilized? With Dooli. So, uh, several pie like, in finance, several Python libraries are commonly used In finance, first, uh, for task, uh, such as the technologies modeling and visualization, here are some notable ones, like Pandas, NumPy, um, and c1, scikit learn, contitlib, tensorflow. So these are the, uh, these are, uh, we can select these are the Python libraries commonly used in finance. And, uh, coming to the utilization. So find out the use for data manipulation and data analytics. Okay. NumPy is essential for numerical compounding, uh, computing. NumPy provides support for large multidimensional areas and matrices as well. Macplotlib and seaborn. Like, these libraries are used for creating static in interactive and statical visualizations. Uh, scikit learn if machine learning is involved, then scikit learn is for tasks such as clarification, regression, clustering. Yeah. And TensorFlow, uh, coming to TensorFlow for deep learning applications in finance, these libraries are widely used.

Discuss real world. So, If you like a real world instance, then, like, uh, if I'm working on a Python based web publication, that, uh, that was explaining varying levels of user traffic. So the application uses database to store And, uh, retrieve the data initially. And the application is hosted on the single server. And as the user grows, it's it it started facing issues during the Speak to system. So to, uh, uh, to address this issue, I decided to leverage cloud architecture principle, specifically using a cloud service like AWS. So, uh, like, a scalability with auto scaling groups such to implement the auto scaling group to automatically address a number of applications servers Installed based on the traffic. Then load balancing. I did the load balancing as well, introduced a a load balancer to evenly, distribute incoming requests among the multiple servers. Uh, after that, I did database scaling, uh, like, uh, the performance So, like, be because the performance bottleneck was in the database, so, like, uh, I used I considered using managed database services like Amazon RDS. So yeah. And, uh, uh, I also explored server

How would you resolve issue with real time Data processing in Python, particularly finance product. So, uh, like, Uh, to resolve issue with the item data processing in, uh, finance product, uh, we can new we can, uh, we can actually be using Python that involves address in performance, reliability, and latency concerns. So, uh, we can use efficient data structures and libraries as well, Like, a level of high performance libraries, like, for efficient data manipulation. Uh, a single we can we can, uh, we can go We can try asynchronous programming as well, like, implement asynchronous programming with libraries like asyncio to handle concurrent task efficiently, uh, then we can, uh, like, we can integrate streaming platforms such as Apache, Kafka, or AWS, Uh, Kinesis for the handling the item data needs. Uh, we, uh, also, we can, uh, like, uh, we can profile and optimize the critical section of the, uh, our code Using tools like c profile or line profiler, we can, uh, we can utilize parallel profile, uh, parallel processing techniques such using libraries as component dot or task. So yeah. And we can also implement the caching mechanism to store and quickly retrieve frequently access data. So yeah. Uh, and after that, like, we can also do the distributed computing. Thing we can also consider, uh, distributed compute distributing our computation across the like, a multiple

How can solid properties effectively implemented In Python? Okay. So, uh, give me, um, give me a second. Let me think about it. So, So the solid principles, like, single, uh, as, like, single responsibility open and close, uh, into it. So, uh, like, Uh, this solid is the are the fundamental design principles of object oriented programming. So implementing them in Python involves applying, like, good software design practices. So, Uh, so we can, uh, we can, like, uh, do this, uh, SRP. Like, that is a single responsibility principle. Well, so ensure, uh, so we can ensure that each class and the model has only one reason to change, uh, separate different concerns to distinct our classes functions. Then after that OCP, uh, like, that is open or close principally, so we can design classes to be, uh, open for extinction but closed for modification. So that after that LSP, that is, uh, list list go substitution principle. So subtypes, uh, must be substitute table for their base types without, uh, altering the current like, uh, sometimes, uh, like, sometimes, we'll substitute for their base, types without altering the correct names of the program. Uh, like, also, we can ensure that the, yeah, the direct classes are under the contracts of their base class. Then we can, uh, ISP interface integration principle client should not be forced, uh, to depend on any interface they they do not use, then the IP was in principle. So depends on the obsession and not, uh, corrections.

So the following, uh, code snippet is missing, uh, something. Assume we are trying to implement a singleton design pattern, what changes would you recommend? And why do you use this design pattern? Okay. Just give me yours. Be quiet as I can to think on it. So to, like, to implement singleton design pattern in Python, uh, we need We need to ensure that, uh, only 1 instance of a class is greater and provided at a global point of access. Uh, Yeah. So and the changes we can make is that we can introduce a class variable, like, instance to source singleton pattern. After that, in the init method, we can check if instance is not already set. Uh, if not, then we can create the instance and check the provided parameters. If an instance already exists, then we can return that existing instance. So now Why, uh, like, now, uh, why, like, why we use single button pattern is, like, purpose of single button is is single instance. So ensure that the class has only 1 instance and provide a global point of access to those that instance. So after that, the glow, uh, we can it is also useful when a single point of control is needed for action that affect the entire system. Uh, in case where we creating multiple instances of a class in is resources is Uh, like, in the case where we're creating multiple instance of class, is resource intensive? Okay. This resource intensive. The single button can help manage or we can say, Uh, the user resource effectively. After that, uh, also, when there is a need to have a single configuration instance Shared across the applications. So for that, we use a singleton pattern. Uh, like, uh, and in the context of, uh, DB connections, my DB connections class, Uh, using a single ensures that there is only 1 database connection instance throughout the application. So, like so preventing Unnecessary resource consumption, and we can select potential issues related to multiple connection. Uh, so, like, Uh, single net method is particularly useful in this scenario where maintaining single point of control for database interaction is

So even the following Python function explain what okay. Even the following even the following by the function, explain what It doesn't and point out any issues you see with it. So, like, the pro info provided Python function, it seems like, uh, attempt of to implement like, you attempted to implement a recursive calculation of the n f Fibonacci number. So, like, however, there are some issue with the code Such as, like, the function calculates the number using recursion. Like, but if n is 1 or 2, it returns 1. So bay so, like, it is a base case. Otherwise, it decursively calculates the Fibonacci number by summing the result of the 2 previous Uh, Fibonacci numbers. Uh, and, like, uh, issues well, The exponential com time complexity is, like, high. So, like, uh, this is this implementation has exponential time complexity. Like, Uh, it recalculates Fibonacci numbers. Then there is a lack of, uh, there is lack of, uh, memo memo We can say there is a lack of memorization or caching mechanism implemented to store previously calculated Fibonacci numbers. So to like, as a result, the function performs redundant calculations. So, like, that will leading that, uh, that will lead to inefficiency. So to address this issue, we can enhance the function by, uh, incorporating the, uh, to solve the intermediate results and reduce the redundant computations whatever here we are giving. So, uh,

What framework do you prefer? Server side logic. How does it? Okay. So what why why don't we do the website logic and why and how does that ensure to the higher responsiveness of a wave application. The the the choice of either framework for server side logic is often depends on the specific requirements of project or, like, team difference and the and the nature of the application. So, like, there are 2 popular mainly there are 2 popular frameworks, uh, for building web applications with Python and and Django and Flash. That is Django and Flash. So, uh, the Django like, if you are going to Django, then advantage. So, like, uh, it is a full featured and follows the batteries included philosophy. So providing a wide range of built in functionality such as ORM, admin panel, Authentication, it is so it is also well documented and follows the MVC that is model you control that. Strong, uh, and it has also strong emphasis on drive. That is don't repeat yourself and convention over confusion. So, uh, and, also, it has a built in security Features and practices. So, like, uh, for ensuring high responsiveness, uh, Django asynchronous, uh, Lee supports allows handling of a large number of concurrency. Flask. So, like, uh, Flask is lightweight and flexible providing essential components for building web applications. So Without imposing a specific structure, more freedom to the for the developers to choose libraries, like and component based on the project model such as the ORM and this. So fact factor for influencing response in the single support and scalability. Yeah. And caching and API. So

What are some scalability challenges you Full seen in this full API, and how would you work around them? So let's get these many challenges in this. Full APIs can arise or as the system grows and handles and handles the increased traffic or data. So, uh, like, Some common challenges and potential solutions are, like, like, uh, challenges are, like, increasing the request load. So, uh, so As the number of the request increase, the EPM may experience a performance de degradation. So so, uh, and we can for to, uh, optimize this, we can implement the load balance and distribute incoming request multiple servers instance. This ensures, like, even distribution of load and prevent a single server from becoming a bottleneck as well. Uh, then also database scalability is also one of the, uh, uh, potent uh, like, uh, one of the challenge. Like, the growing user base can lead to increase in database and load potential and performance as well. So, like, you so, Uh, to optimize this, we can utilize database sharing or replication as or caching mechanism as well. After that, the latency and, uh, the like, uh, the as API handles more requests, the user one time may increase impacting user experience. So To optimize database queries, you, uh, we can use caching on frequently accessed data, uh, and implementing in the content, Uh, delivery network as well. Uh, maintaining and also there is a challenge of maintaining state in scalable manner, uh, can be challenging as well. So design, uh, we can design the API to be stateless where the possible, uh, like, we can use use stateful components, like database for storing