Kurapati Venkata Krishna Gopinadh

Hi, my name is Gopinath, and I'm currently working in Cognizant Technology Solutions as a data analyst, where I deal with clients from different countries other than North America, where we deal with a banking and insurance company, handling data related to banking. We create visualisation charts using Power BI and will do data modelling using SQL servers. I completed my M.Tech in BITS Pilani in October 2023, and I also hold a BTEC degree from Lawley Professional University, earned in 2020. I'm proficient in ML, Python, AI, NLP, and deep learning, and I have good knowledge on Power BI and Power BI technologies, as well as SQL server technologies.

Chatbots' codebase principles include needing to provide an accurate result based on the user's prompt, offering a very active response according to the current situation asked by the user, and responding immediately. There should be no repetition, such as multiple repetitions, which can be irritating to the user. The chatbot should provide the appropriate solutions and make the user experience flexible. If the chatbot gives good suggestions, it will be beneficial, and these are the solid principles.

how to optimize a SQL query that aggregates data across multiple tables for a chartboard response basically we can use different techniques in a SQL query optimizations we can use normalization techniques first normal second normal and third normalization SQL query we can use like multiple joins and we can aggregate we can use aggregate functions and we can use CTEs combined expressions and to join multiple tables and we can create procedures or views to aggregate data we can add for aggregating we can use aggregate functions like sum count and minimum we can use group by aggregate functions where using group by we can count or we can sum up the columns in this way we can aggregate the data from multiple tables basically we can use join basically we can join the tables and after that we can aggregate using group by statement where the aggregate functions are sum minimum maximum and count these are the aggregate functions we can use in the SQL query.

To check the chatbot performance, we basically need to check the accuracy and then how it is, a chatbot response is mainly we depend on NLP, natural language processing techniques, where we need to take the sentence, divide it into vectors, prompt the model, send the vectorized string to the model, and predict the required output. For better performance, we can use a sentence analyzer and metrics like precision to check if the required solution is being given. We can also check for optimal solutions by using the latest technologies for the chatbot. The metrics we can use for performance check are these techniques.

We can handle concurrency using a, actually, by using different cloud technologies and maintaining load balancing and each interacting session in different clusters so that it does not match with the other conversations. So, using these cluster technologies or server load balancing technologies, we can handle concurrency in the chatbot while using multiple users. We can use cloud services to do load balancing for the chatbot, which maintains separate chats for each user, so it will not become difficult for the users.

We can use different techniques for handling overfitting like regularization techniques, so we can use regularization techniques and we can use multiple optimization optimizers, such as the Adam optimizer, according to the data and the model used, to better prevent the overfitting of a model. We can also use different standardization techniques, standardizing the data so that it maintains consistency across all rows in terms of mean, median, and mode, in the standardization mode. Data will be arranged in the standardization mode to prevent overfitting. We can also check by different learning modes or optimizer techniques for this neural network model to prevent overfitting.

page data query and database query in database, if query in record data results append I got okay results list of dictionary items with data as a key so basically here why performance issue is it will check each database and it will check whether the database is present or not and whether the record table is present or not and after that only it will query on the tables so it will take time to check all the databases and all the records which are present in the database so it will be a performance issue so the chatbot will respond slowly to the user for this code.

here basically, here we are doing this, it is taking a word's list and after it is finding the 'ing' in the ending of each word, if the 'ing' is present in the word, it is, it is, it is taking the substring, it is deleting the 'ing' and after that it is adding in an ArrayList so that and it is returning the ArrayList, here basically, it created a function, public function, stem words, it is, in the ArrayList creation, in that, the list was not mentioned, may be this caused a mistake in the compilation error and if you see, the for loop is fine, it is a word is a string and it is a substring, taking zero from word length, it is taking minus 3, here mistake is, it will, it will, the list contains only the words which are removed, which are 'ing' removed, it will not contain the, all the words, this may be, this may be one of the mistakes because all the words need to be present in the list but the word need to, but the word contains 'ing', it need to be removed, I think this is the logic, this is the code they want to develop, may be this, that is one of the logic bugs in this code, where it adds only the 'ing' words which are removed, which are removed words only added in the list, all the words will not be added in the list.

for catching strategy we can use different catching methods basically I am not much aware of these catching strategies so where we have different catching technologies where it stores different prompts when we say the cloud we can store different prompts so that when the user response given these type of prompts it automatically gives the response according to the prompt these can be cloud-based so in this way we can use it to implement a catching strategy these are the things I know about catching strategies.

Sentiment Analysis is used to determine whether a person is in a good or bad mood, or in a positive or negative mood. We will find the sentence, vectorize it, and then identify any negative words, such as keywords, "not," "is," any negative things, "no," and words that reflect negative sentiments, like these negative implemented keywords present in the sentence. We will think of it as a negative sentence, indicating the user is in a bad mood, so we can respond with positive sentences to make him feel better. We can use Sentiment Analysis to vectorize and find keywords, and if the positive keywords are more, the percentage of the positive will be high, and if the negative keywords are more, the percentage will be high. If the negative percentage is high, we can tell the sentence is a negative sentiment, and we can respond positively to make the user interactive and cool. If the user is in a positive mood, we can make him laugh and make him feel comfortable, and he can chat more to the chatbot and get the result he requires.

The importance of continuous integration and deployment is very much important because nowadays technologies are increasing, so models are also increasing, and user requirements are also increasing. We need to find whether any new bugs are coming into the code, automatically resolve queries, continuously integrate new updates, and deploy new code in the chatbot, so it can easily respond to the user without any lag and delay in response. Continuous integration and deployment is very much important because the chatbot interacts with the user according to the query, so it needs to be in a continuously monitoring mode where if any errors occur, we need to resolve them immediately within 5 to 6 hours. Integration and deployment is very much important in a chatbot.