Sri Megha V.

🚀 Just completed the "Building and Evaluating Advanced RAG!" course! 🚀 This journey deepened my understanding of Retrieval-Augmented Generation (RAG) techniques, with hands-on exploration of Direct Query Engine, Sentence Window Query Engine, and Auto Merging Query Engine. I also gained valuable insights into LLM evaluation with tools like LLama Index and TruEra, learning how to assess model performance through key metrics such as latency, groundness, answer awareness, context relevance, and total cost. Excited to bring these new skills into practice and contribute more effectively in the rapidly evolving AI landscape! 🌟 #AI #RAG #LLM #MachineLearning #ArtificialIntelligence #DataScience

7

This a foundation course if you want to land up into the world of forecasting. I have done few forecasting projects earlier, but this course has exposed me to some lesser known fact essential if you want be ace time series modeling.

7

Har Har Mahadev! 🙏🙏🙏 thanks to Krish Naik iNeuron.ai Overcoming Categorical Data Issue in Data Science: A Case for SMOTE Real-life data is often plagued with a class imbalance problem, a common problem that is predominant in fraud detection, medical diagnosis, churn prediction, among other domains. Model performance suffers considerably due to the fact that machine learning algorithms tend to be biased toward the majority class. What are the Ways to Solve this? Fortunately, there is a method called SMOTE (Synthetic Minority Over-sampling Technique). SMOTE creates new minority class observations that allow the distribution of classes within the data set to be more even, resulting in a better predictive model. The way SMOTE works is as follows: Creation of Synthetic Data: Despite being a very concentrated approach, the SMOTE technique eschews blind duplication of existing data where it interpolates the space between two (or more) samples from the minority class to generate new observations. Enhanced Model Effectiveness: After the distribution of classes has been addressed, classifiers such as Logistic Regression and decision trees will be able to learn more effectively how to separate the classes. Implementation of SMOTE: It is straightforward to use with packages such as imbalanced-learn for Python: How to Apply SMOTE: It’s easy to implement using libraries like imbalanced-learn in Python: from imblearn.over_sampling import SMOTE smote = SMOTE() X_resampled, y_resampled = smote.fit_resample(X, y) Key Points: SMOTE works best when applied before training. Combine it with techniques like Random Undersampling for even better results (SMOTEENN). Use performance metrics like precision, recall, F1-score, and AUC-ROC to evaluate model performance on imbalanced data. Balancing your dataset with SMOTE can make a significant difference in the accuracy and fairness of your predictions! #DataScience #MachineLearning #ImbalancedData #SMOTE #AI #DataScienceTips

2

Hey everyone! Exciting news—I was invited to join a YouTube podcast on Generative AI with Learnomate Technologies Pvt Ltd Please check out the video, give it a like, drop a comment, and share it as much as you can. Hope you all enjoy it! ❤️ #Datascience #GenerativeAI #Podcast #AIInnovation #LearnomateTechnologies #ArtificialIntelligence #MachineLearning #Deeplearning #AICommunity #TechPodcast #AIInsights #FutureOfAI #LinkedInCommunity #TechTalks #AIContent #DataScience #AITrends #Innovation

55

2 Comments

Higher-Order Functions (TRANSFORM and FILTER) are now Generally Available in Snowflake. In my latest blog, I delve into how Snowflake empowers users to perform advanced transformations with ease, making complex data analysis a breeze on the cloud data platform. Keep pushing boundaries and embracing the future of data management with Snowflake! Link to the Blog in Comments section. #Snowflake #SolutionInnovation #DataTransformation #DataCloud

28

1 Comment

After reading the paper on Chatterjee's Xi correlation coefficient, I was eager to learn more and see it in action. Taking the time to read through Chatterjee's original paper with my good friend Mark Stent really helped my understanding, but also helped me clarify and readjust how I should be looking at this new approach. This is not here to replace traditional correlation coefficients, but in fact can be used in conjunction with traditional methods for improved results. Obviously, I was eager to see this in action, and thankfully my buddy Mark came through again by sending across this shiny new research paper. The results seem promising and the overall idea of combining Spearman's and Chatterjee's for a test of independence is very easy to digest. I look forward to seeing whether the applications of these new methods play out successfully in my own work. https://lnkd.in/eM8p76cZ

8

1 Comment

Thrilled to announce that I've completed the Generative AI Fundamentals course by Databricks! Big thanks to Tredence Inc. for supporting my learning journey. Excited to apply these new skills and drive innovation! #GenerativeAI #AI #Tredence #Databricks #ContinuousLearning

24

4 Comments

🍁 Thankful for Innovation and Learning this Thanksgiving! 🍁 I’m incredibly grateful to Snowflake for the opportunity to attend BUILD 2024 and for this wonderful swag! The event was packed with cutting-edge sessions on Machine Learning, NLP, and Generative AI, showcasing how Snowflake continues to innovate in the AI and data space. 💡 Key Highlights from ML, NLP, and GenAI Sessions: 1️⃣ Fine-Tuning LLMs with Snowflake Cortex Learned how to customize LLMs for specific tasks, reducing computational demands while maximizing accuracy. Fine-tuning techniques are opening new doors for business-specific applications. 2️⃣ Conversational Applications and NLP with Cortex AI Cortex AI is redefining how we interact with data, from operationalizing NLP pipelines to building conversational apps with advanced text-to-SQL enhancements. 3️⃣ End-to-End ML with GPUs in Snowflake Notebooks Witnessed the seamless development and deployment of resource-intensive ML workflows like deep learning, all within Snowflake using Snowpark. 4️⃣ Building Enterprise-Grade AI Apps with Snowflake Native Apps Explored how AI-powered predictive applications integrate with Snowflake's ecosystem to deliver high-impact insights while ensuring data security. 5️⃣ Evaluating LLM Applications at Scale Deep dived into techniques for improving LLM reliability and addressing challenges like hallucinations and safety issues. The practical demo on the RAG triad framework (context relevance, groundedness, and answer relevance) was particularly inspiring. 6️⃣ Building RAG Chatbots Using Cortex AI Hands-on session to build a Retrieval-Augmented Generation (RAG) chatbot combining structured and unstructured data to deliver accurate, real-time answers. 💡 Key Takeaway: Snowflake is making it easier than ever to build, scale, and productionize ML and AI applications, empowering developers to leverage GenAI and advanced NLP tools to unlock the true potential of their data. 🎉 A huge shout-out to Sridhar Ramaswamy, Andrew Ng, and all the brilliant speakers for their inspiring sessions. Your insights are shaping the future of AI-driven innovation! This Thanksgiving, I’m thankful for the opportunity to learn, grow, and be part of a community dedicated to driving data and AI forward. #Thanksgiving #SnowflakeBUILD2024 #MachineLearning #GenerativeAI #NLP #AIInnovation #DataScience

74

Hello connections, I see a many of aspiring data scientists do practice RAG but it Its is much deeper than just to retrieval of closest similar chunk from vectorDB and pass to LLM to generate responses. I have started going deeper into LangChain and RAG... yet another step towards scalable AI chat bot to answer the questions by looking into the relevant documents for the reference and generating the best possible responses to answer. #LangChain #Deeplearning #rag #LLM #Chatbot #ai #datascience

9

2 Comments

I had the incredible opportunity and honour to speak at the second session of the AI/ML Talk Series at the Indian Institute of Information Technology, SriCity, where we delved into some critical topics in machine learning, such as how to correctly handle imbalanced classes in classification, cost-sensitive learning, feature interaction and monotonic constraints, probability calibration, and an introduction to conformal prediction. A heartfelt thank you to Dr. Rakesh Sanodiya for inviting me and Indian Institute of Information Technology, SriCity for hosting me and to everyone who attended! A huge shoutout to the amazing team for organizing this session, and thank you for all the enthusiasm and engagement! 🎉 I’m glad to see how well these ideas resonated with everyone. Let’s keep the conversation going! 💬 Feel free to connect with me if you want to dive deeper into these topics or share your insights. Together, we continue to grow and push the boundaries of what AI/ML can achieve. 💡 #MachineLearning #AI #DataScience #ConformalPrediction #ExplainableAI #XAI #DemystifyingML Links to some of my posts discussing these topics: Probability Calibration - What, Why, How?: https://lnkd.in/dtaiU7NB https://lnkd.in/ddGyq7BP Don't use SMOTE, use these techniques instead to handle class imbalance: https://lnkd.in/dBCrJB3k Feature interaction and monotonic constraints in Ensemble models: https://lnkd.in/dKimuVeN

20

In light of the recently concluded Andhra Pradesh Assembly elections and their results, I'm excited to share my new Tableau dashboard! 🎉 This dynamic tool offers: 📊 Key Metrics: Analyze victory margin, votes received by the winner, party performance, and a constituency-wise breakdown map. 📈 Trend Analysis: Compare past election results over time using an Year filter. (Data used only since 2014) 🗺️ Geographical Insights: Detailed maps showcasing regional variations. Hover over each constituency to see winner votes and victory margins. 🔍 Year Filter: Customize your view by filtering data by year. Designed for analysts, political enthusiasts, and researchers, this user-friendly dashboard provides deep insights into Andhra Pradesh's electoral landscape. Explore the dashboard and share your feedback! #DataAnalytics #Tableau #Elections #AndhraPradesh #DataVisualization #PoliticalAnalysis https://lnkd.in/gSTU-5mN

8

𝗧𝗼𝗽 𝗬𝗼𝘂𝗧𝘂𝗯𝗲 𝗖𝗵𝗮𝗻𝗻𝗲𝗹𝘀 𝗧𝗼 𝗟𝗲𝗮𝗿𝗻 𝗗𝗮𝘁𝗮 𝗔𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀 𝗙𝗼𝗿 𝗙𝗿𝗲𝗲 📊 𝗞𝗵𝗮𝗻 𝗔𝗰𝗮𝗱𝗲𝗺𝘆 While not solely focused on data analytics, Khan Academy offers foundational knowledge in statistics and probability, which are crucial for understanding data analytics. ↳ https://lnkd.in/g5A-jV4s --- 𝗞𝗿𝗶𝘀𝗵 𝗡𝗮𝗶𝗸 Krish Naik’s channel covers a wide range of data science topics, including machine learning, data visualization, and deep learning. ↳ https://lnkd.in/gWxnxRp8 --- 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗠𝗼𝘀𝗵 This channel is an excellent resource for learning programming languages and frameworks, which are crucial for data analytics. ↳ https://lnkd.in/gEiiggAF --- 𝗦𝘁𝗮𝘁𝗤𝘂𝗲𝘀𝘁 StatQuest offers detailed explanations of statistical concepts and data science techniques. ↳ youtube.com/c/joshstarmer --- 𝗙𝗿𝗲𝗲𝗖𝗼𝗱𝗲𝗖𝗮𝗺𝗽 FreeCodeCamp provides extensive tutorials on programming and data science, including Python, data visualization, and machine learning. ↳ youtube.com/c/Freecodecamp --- 𝗠𝗜𝗧 𝗢𝗽𝗲𝗻𝗖𝗼𝘂𝗿𝘀𝗲𝗪𝗮𝗿𝗲 MIT's OpenCourseWare offers free lectures from MIT courses, including data science, machine learning, and artificial intelligence. ↳ youtube.com/c/MIT --- 𝗦𝘁𝗮𝗻𝗳𝗼𝗿𝗱 𝗢𝗻𝗹𝗶𝗻𝗲 Stanford Online provides access to lectures and seminars from Stanford University, covering data science and related fields. ↳ https://lnkd.in/gBMc7XJn --- 𝗖𝗼𝗿𝗲𝘆 𝗦𝗰𝗵𝗮𝗳𝗲𝗿 Corey Schafer’s channel is known for its in-depth tutorials on Python, which is essential for data analytics. . ↳ youtube.com/user/schafer5 --- 𝗗𝗮𝘁𝗮 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 𝗗𝗼𝗷𝗼 This channel offers comprehensive tutorials on data analytics, machine learning, and data visualization. ↳ https://lnkd.in/gnCsMVds --- 𝗦𝗶𝗺𝗽𝗹𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 Simple Learning focuses on data science and analytics topics, with clear and concise explanations. ↳ https://lnkd.in/gVyxTBvW --- 𝗞𝗮𝗴𝗴𝗹𝗲 Known for its competitions and datasets, Kaggle's YouTube channel also features tutorials and lectures on data analytics techniques and tools. ↳ youtube.com/c/Kaggle --- 𝗔𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀 𝗩𝗶𝗱𝗵𝘆𝗮 This channel provides tutorials on data analytics, machine learning, and artificial intelligence. It's a great place to start for practical knowledge and real-world applications. ↳ https://lnkd.in/gEHX4acb --- 𝗖𝗵𝗮𝗻𝗱𝗼𝗼 Chandoo is an excellent resource for mastering Excel, a vital tool for data analysts. ↳ youtube.com/user/chandoo --- 𝗘𝘅𝗰𝗲𝗹 𝗜𝘀 𝗙𝘂𝗻 This channel offers comprehensive Excel tutorials, focusing on data analysis, formulas, and visualization. ↳ https://lnkd.in/djyRabMY

5

𝑻𝒊𝒑𝒔 𝒇𝒐𝒓 𝒏𝒆𝒘 𝑫𝒂𝒕𝒂 𝑺𝒄𝒊𝒆𝒏𝒄𝒆 𝑴𝒂𝒏𝒂𝒈𝒆𝒓𝒔 I became a manager in data science early in my career. It has been a journey fraught with failures and some successes. Here are some of my learnings along the way - 📈 𝑹𝒐𝒍𝒆, 𝒓𝒆𝒔𝒑𝒐𝒏𝒔𝒊𝒃𝒊𝒍𝒊𝒕𝒚 𝒂𝒏𝒅 𝒑𝒓𝒊𝒐𝒓𝒊𝒕𝒊𝒆𝒔 Early on, my vision was small which narrowed the value I was able to create. I had two focuses - 1) ship out high quality work, 2) coach & promote my team. After >10 years in leadership roles, here is how I NOW think of the responsibilities of a data science manager in the hierarchy of importance - 🔹 Build authentic relationships with stakeholders and team members 🔹 Identify and align on highest impact work the team can enable and unlock 🔹 Set up systems, processes, and people to execute on these projects 🔹 Manage projects and priorities 🔹 Coach your direct reports on execution Depending on the seniority of the role, the distribution across these may vary but the core buckets remain the same. 🏆 𝑩𝒆 𝒊𝒏 𝒕𝒉𝒆 𝑨𝒓𝒆𝒏𝒂 Don’t be carried away by the business of meetings that make you feel important but don’t create real value. Spend at least 1/3rd of your time towards hands-on work. Don’t get rusty on the technical chops. Be an SME in several areas in the workflow - tech stack/tools, data models, methodologies, algorithms and specific tables. This will build trust with your reports while reducing dependency, increasing team efficiency and productivity. 🔄 𝑻𝒂𝒊𝒍𝒐𝒓 𝑷𝒓𝒐𝒄𝒆𝒔𝒔𝒆𝒔 & 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒔 Data science lacks standard processes, unlike software engineering, due to its evolving role, inherent variability in methods depending on project goals. Creating processes and systems that will help with the ground reality of the ecosystem and culture at your company, is important. 🎲 𝑼𝒏𝒅𝒆𝒓-𝑷𝒓𝒐𝒎𝒊𝒔𝒆 𝒂𝒏𝒅 𝑶𝒗𝒆𝒓-𝑫𝒆𝒍𝒊𝒗𝒆𝒓 In the unpredictable world of predictive modeling 😉, where factors such as data anomalies, algorithm biases, and unexpected market shifts can disrupt the accuracy of forecasts, it pays to temper expectations and try to exceed them. #datascience #analytics #businessanalytics #leadership #techleadership #careers #techmanagers

28

1 Comment

🚀 Looking for Clients Ready to Innovate with AI, Data Science, and Advanced Software Solutions! 🚀 Are you a business or organization looking to harness the power of machine learning, NLP, data engineering, or backend development to drive impactful results? Let’s work together to turn your ideas into reality! With extensive experience in creating custom solutions—from synthetic datasets and data pipelines to language model fine-tuning and backend development—I can help bring advanced technical expertise to your project. My work is focused on delivering high-quality, scalable, and efficient solutions tailored to meet your needs. 🔗 For more about my expertise and client success stories, check out my Upwork profile: https://lnkd.in/gipwtHBB If you’re ready to take your project to the next level, please feel free to reach out. Let's transform your vision into actionable, data-driven solutions! #AI #DataScience #MachineLearning #NLP #SoftwareDevelopment #Upwork #ClientProjects #Innovate

24

Very common rookie mistake candidates do while trying to get a job in the data science industry. Having the following 3 projects from #Kaggle in your resume: House Price Prediction Telco Churn Prediction Titanic Problem No, you will not get your resume selected with these projects. I get it. YouTube tutorials suggested doing these projects so you are wondering, “Why aren’t my resume getting picked after I’ve completed them?” Well think about it. As a hiring manager, how do they know you didn’t just copy an existing notebook? You might say, “But they can at least give me a chance to prove my skills!” Again think from their perspectives - What if they do give a chance but you don’t perform up to the mark, that’s valuable time wasted. Now imagine giving that time to all candidates. To really get noticed, work on projects that will show that you’re not just another candidate. And there are ways. Many ways. We will discuss some during https://lnkd.in/gMNckgRi on 14th October, 9pm. #DataScience #DataAnalytics #HiringTips #CareerAdvice #DataJobs #Kaggle #ResumeTips #StandOut #DataProjects #JobSearch #Job

9

🏏 Data Analytics Transforms Cricket Insights: A Game-Changer for Top Players! 📊 In the world of cricket, data analytics is revolutionizing how players and teams understand and improve their game. By leveraging advanced analytics, we’re witnessing unprecedented insights into player performance and strategy. 🔍 Spotlight on Key Players: Rohit Sharma: Dive into how Sharma’s striking patterns and performance metrics are reshaping batting strategies. Jos Buttler: Explore how Buttler’s aggressive playstyle is analyzed to optimize match-winning strategies. Mark Wood: Understand how Wood’s bowling data is fine-tuned for maximum impact and precision. David Malan: See how Malan’s scoring trends and shot selection are being used to enhance batting approaches. The integration of data analytics into cricket not only enhances individual player performance but also drives strategic decisions on the field, leading to more dynamic and competitive games. 📈 Key Benefits of Data Analytics in Cricket: Improved player performance through detailed statistical analysis. Enhanced strategic planning and game tactics. Personalized training programs tailored to each player's strengths and weaknesses. Cricket is evolving, and data analytics is at the forefront of this transformation. It's an exciting time for fans and analysts alike as we witness the synergy between technology and sports! #Cricket #DataAnalytics #SportsScience #RohitSharma #JosButtler #MarkWood #DavidMalan #SportsAnalytics #PerformanceOptimization #CricketInsights Please find this video :- https://lnkd.in/db5BEFSz Don't forget to Like & Subscribe Datawithkhan my YouTube Channel for more information and Knowledge.....

15

Thrilled to announce that i have cleared Databricks certified #Generative #AI Engineering Associate certification. The exam assess you on fundamentals of LLM, LLMOps, Langchain, RAG, Governance, Model Evaluation. Thankful to Databricks Academy to provide me with appropriate resources to prepare for this exam! Look forward to next one! #databrickslearning

73

17 Comments

Hi Connections, I’m super excited to launch a new series dedicated to two-way knowledge sharing with all of you amazing folks! This series is all about learning and growing together, benefiting everyone involved. 📊 Day 1: Have you ever tried Hive partitioning in Databricks? 🐝 Question: How does Delta Lake handle partitioning differently? Answer: The Delta table consists of Parquet files structured in nested directories, with a transaction log containing information about the files and partition structure. When engines query the Delta table, they use the transaction log to figure out file locations and partition information, avoiding the need for file listing operations. Interestingly, physical partitions in a Delta table are unnecessary and could be logically partitioned. Delta Lake supports physical partitioning mainly for compatibility with other engines that use Hive-style partitioning and to facilitate conversions. Let’s make this post more interactive! Have you tried Hive partitioning in your project when using Lakehouse architecture? Share your experiences and insights in the comments below! 💬 Feel free to tweak it as needed! 😊

69

4 Comments

The Buzz word every where is GenAI , LLM's and Multi models , lets discuss on use cases where its driven significant value and not just made life slightly easier for stake holders. Comment below and lets discuss. #genai #llm #datascience

6

6 Comments