Esha Punjabi

With Sarath Shekkizhar, we are exploring the geometric structures behind LLM reasoning capabilities. So far, our findings suggest: 1. Higher density in self-attention graphs implies higher intrinsic dimension. 2. Higher intrinsic dimension induces more expressive power/reasoning with MLPs. 3. Longer context and more attention heads can boost intrinsic dimension. This geometric lens offers new insights into LLM performance and potential improvement strategies. Early days, but we believe it is a promising direction! Stay tuned for more as we dive deeper into the geometry of AI reasoning. Read our pre-print: https://lnkd.in/gcekSWYD #AI #MachineLearning #LLM

25

I remember when this all started. Faris - invited me to catch up at our favorite spot around Blok M. I had no idea he’d just wrapped up a procedure at Pondok Indah Hospital! (He really knows how to add drama to our catch-ups.) We’d been on separate paths and projects since "the winding down project", but there he was, reminding me of our dream to ride this AI wave together again. It was a plan we’d hatched with Sahal Zain & Gurnoor Dhillon when the world of AI was still mostly theoretical and jargon-heavy. Somehow, the conversation continued with Gurnoor Dhillon while he was waiting for his meeting with someone who would eventually become one of CalvinBall’s key early adopters. It felt like everything was coming together, like a series of seemingly random dots connecting to form a bigger picture. Seeing how far we’ve come as a team at CalvinBall Technologies is truly exciting. What started as theoretical discussions has now evolved into practical Gen AI solutions that are solving everyday problems and hold endless possibilities to shape the industry. 👨‍🚀 👩‍🚀 Here’s to pushing boundaries and achieving new milestones together! Gurnoor Dhillon, Sandeep Ramesh, Faris -, Sahal Zain, Eva Zuliya, Ratu Nadiah Khairunnisa 🚀 #GenAI #AIInnovation #TechJourney #ArtificialIntelligence #TechTransformation #CalvinballTech #Calvinballers

55

2 Comments

⏫ Leveled up my understanding of retrieval systems (R+AG) 🔍 and put those learnings to work by improving GitHub Issues management! Here’s a breakdown of what I explored: ✨ Embedding-based Retrievals – How embeddings drive more meaningful search by mapping concepts into vector spaces. ⚠️ Pitfalls of Simple Vector Search – Identified common issues like irrelevant results and learned how to tackle them effectively. 🔍 Query Expansion with LLMs – Used Large Language Models to broaden search coverage and uncover deeper insights. ⚖️ Cross-Encoder Re-Ranking – Tuned result rankings with cross-encoders to bring the best matches to the top. 🛠️ Embedding Adaptors – Learned how to tweak embeddings for different tasks to keep retrieval systems flexible and efficient. Super excited to apply these concepts further! If you're working on retrieval systems, let’s connect and exchange ideas! 🚀 #AI #ChromaDB #Retrieval #LLMs #GitHub

15

5 Comments

🛑 𝐌𝐚𝐬𝐭𝐞𝐫𝐢𝐧𝐠 𝐃𝐚𝐭𝐚 𝐒𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐬 & 𝐀𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦𝐬 (𝐃𝐒𝐀) 𝐚𝐥𝐨𝐧𝐞 𝐰𝐨𝐧’𝐭 𝐡𝐞𝐥𝐩 𝐲𝐨𝐮 𝐚𝐜𝐞 𝐃𝐒𝐀 𝐢𝐧𝐭𝐞𝐫𝐯𝐢𝐞𝐰𝐬! Yes, technical skills are important. But, your interviewer is also assessing something equally critical: 𝐡𝐨𝐰 𝐰𝐞𝐥𝐥 𝐲𝐨𝐮’𝐥𝐥 𝐰𝐨𝐫𝐤 𝐰𝐢𝐭𝐡 𝐭𝐡𝐞𝐦 𝐚𝐧𝐝 𝐭𝐡𝐞 𝐭𝐞𝐚𝐦. 💼👥 It’s not just about solving problems. It’s about how you communicate, collaborate, and approach challenges under pressure. 🔑 𝐄𝐱𝐚𝐦𝐩𝐥𝐞: How effectively do you explain your thought process? In this video, I dive into how DSA interviews test your communication skills 📹👇 🎥Check out the full video series on technical interviews to learn how to ace the entire process: https://lnkd.in/g3U6jvzU #TechInterviews #CommunicationSkills #DSA #CareerTips #InterviewPrep #SoftSkills #TechnicalInterviews #WorkCulture #JobHunt #CareerGrowth

14

1 Comment

Looking for recommendations for good hindi-english (hinglish) code-mixed speech to text models/corpuses. Anyone building anything with speech in the Indian context probably realized normal ASR systems fall flat, losing >50% context because of their inability to handle hinglish conversation (which is extremely common for a huge chunk of modern India). See the title of the first video when looking for shark tank. Once you start noticing it, its hard to unnotice. If speech to text doesn't work, every downstream task thas cascading failures. I'm not looking to build anything commercially so hopefully #research corpuses will also be okay. Literally only want to progress speech for India and this is a major blocker. This is one of those areas where other countries can't innovate for us and hand us a solution unfortunately. #speech #asr #codemix #hinglish #stt

3

1 Comment

🚀 Making LLMs work for you - A Guide to Become 10x Programmer/MLE/DS I am sure many of you use LLMs for your day-to-day work. Mostly, I have seen people use LLMs while writing code (Copilot), or as an alternative to Copilot to catch bugs, but they have become much more useful than that, provided you know how to use them well. Here are a few ways I have been using LLMs in my daily work that are helping me increase my productivity: 📚 1. Quick Paper Assessment with ELI5 Summaries - Paper abstracts are called abstracts for a reason. Asking an LLM for an ELI5 summary makes it much easier to know if I want to read a paper. 💻 2. Implementing Research Papers in Code - Try giving an LLM like Claude/ChatGPT a research paper and ask to code it. You would be surprised. Tip: Claude performs much better for this task. 🔄 3. Upgrading Code Structure and Readability - Do you want to add typing hints when writing your Python code, or do you need comments in your code, or do you have a big function that you want to be converted to classes in your code? LLMs can do it all, and you would be amazed at the code-conversion power they have. ⚡ 4. Optimizing Code Performance - The thing with LLMs is that if you tell them what to do, they can do it well. For example, if you give them a non-optimized piece of code that runs correctly, they will be able to write it in a much-optimized way. For example, I just used an LLM to convert a matrix multiplication in PyTorch to sparse MM as I was getting an OOM memory error because one of the matrices became too large. I am not even joking, but that would have taken at least 1-2 hours for me to come up with. 📝 5. Automated Documentation - Got bored of writing documentation for what you did but have the code? LLMs also work the other way around. Want to write a summary of what you did in a week, you can use an LLM for that. Saves so much time that gets wasted in formatting. 🧪 6. Streamlined Test Framework Creation - Let LLMs handle the boilerplate for Unit tests & Integration tests while you focus on core logic. 🌐 7. Language-Agnostic Development - You can paste a function in Python and ask for its equivalent in Rust, complete with language-specific best practices and error handling. This dramatically reduces the learning curve when working with new languages and helps maintain productivity during language transitions. 🤝 Apart from all of the above, LLMs work well as mentors as well as interns. LLMs are great when you just need someone to bounce ideas with. I sometimes just talk to an LLM as I would to a colleague talking about the problems I have and how I can solve them. And they come up with the right ideas or send me in the right direction a lot of times. They also act as a personal assistant for free, who can deal with writing messages, creating charts, and doing all sorts of uninteresting work for you. 💭 What's your favorite way to use LLMs? Let me know in the comments! #MachineLearning #AI #Programming

47

🚀 In today’s Google earnings call, Sundar Pichai shared a striking stat: *25% of code going to production is now AI-generated*. This got me thinking—how is this reshaping the role of Software developers? A friend recently pointed out that they’re hiring fewer junior engineers, as LLMs now handle tasks like writing test cases and creating small, modular pieces of code. It seems we may see a shift toward more specialized engineering roles, where highly skilled developers are increasingly in demand. Companies now value an engineering mindset(focused on solving real problems) over just coding skills—developing it is challenging.

75

13 Comments

Over last few months, I am increasingly leveraging LLMs to write code. From 10% to nearly 78% currently (based on lines of code) of my code is from LLMs. I have my own pipeline to fine tuned LLMs to augment my needs. Observations - 1. I am producing more code --> MORE IMPACT (approx 2.3X) 2. I am reading more code --> MORE MAINTAINABLE CODE 3. I am learning more as now I can start from concept and expand out ---> MORE LEARNING 4. I am incorporating best practices --> BETTER CODE 5. Lesser typing --> LOWER REPETITIVE STRAIN INJURY #AIBUDDY

76

8 Comments

Prof. Rao's post is a good starting point for LLM practitioners to become familiar with classic (i.e., predating LLMs) AI concepts and theories, for two reasons: 📌 To understand that some things are theoretically beyond the reach of LLMs and not waste resources chasing them. 📌 To hedge against the possibility that LLMs may be replaced by other technologies and/or undergo fundamental technological revamps, which would render all existing investments in LLMs obsolete. #artificialintelligence #machinelearning #deeplearning

41

3 Comments

Last week at HackMIT, my team (Sanskriti Akhoury, Krish Modi, and Aly Shariff) and I built Runk: an AI-powered e-commerce platform that understands your preferences to display relevant reviews and products. How do we know your taste? Using autoencoders, we assign each user a profile embedding, which our own model updates as you leave more reviews. With these profile embeddings, we rearrange e-commerce search results to display products with the most positive reviews left by people most similar to you first. We also built a Google Chrome extension to interface Runk onto pre-existing sites like Amazon to re-rank these reviews based profile similarity. A huge thank you to HackMIT for making this hackathon an amazing experience. I also enjoyed meeting new people at Massachusetts Institute of Technology, Harvard University and Northeastern University! 👾https://lnkd.in/eyhU3nvK Outside of this hackathon, here’s what I’ve been up to this summer: • Built Guardian at HackThe6ix - an AI wearable for SOS situations. With a Raspberry Pi, Gemini Flash, and Hume AI, it contextually analyzes audio and video of emergencies and sends a summary to first responders. Collaborators: Sanskriti Akhoury, Daniel Hong, and Brian Zhang. 👾https://lnkd.in/ei5-Gu46 • Built Siren - an AI-powered dashcam. It analyzes red light camera runs, illegal turns, speeding, and other driving behaviors using YOLOv8 and algorithms built with reverse geocoding. 👾https://lnkd.in/eKMs2x5t • Joined Mimrr as a Machine Learning Engineer to work on their LLMs for automating code documentation. • Joined WATonomous as an Autonomous Software Engineer to work on self-driving cars. You can view these projects on my new website! 👾https://lnkd.in/e-hWsAxB

987

20 Comments

🚀 Excited to share that I tackled a medium-level LeetCode problem today! 🎉 The challenge was to return all numbers in the range [1, n] sorted in lexicographical order, with an O(n) time complexity and O(1) extra space. It was a great exercise in algorithmic thinking and efficient coding in Java. Always rewarding to push my problem-solving skills! 💻✨ #LeetCode #Java #CodingChallenge #Algorithm #ProblemSolving

25

2 Comments

🚀 Excited to introduce our latest blog "The LLM Trifecta: Balancing Cost and Performance with Accuracy for a Real-World Insurance Copilot". ✍ Co-authored by Hitanshu Machhi, Shreya Dikshit, and Abhishek More 🔍 LLM-powered chatbots are revolutionizing customer support in the insurance industry. In this blog, we share our insights on reducing costs while maintaining top-notch performance and accuracy. 💡 Learn about: 1. The factors driving LLM costs and how to optimize them. 2. Challenges we encountered and the innovative solutions we devised. 3. Strategies for cost reduction and optimization, including data preprocessing, custom agent development, and multi-model processing. 4. Real-world results: a 20-fold reduction in LLM usage spend and significant improvements in response times and accuracy. 🔗 Read the full blog post here: https://lnkd.in/g5CTs2vv | Mindstix Software Labs 👏 Special thanks to Saloni Luktuke, Deeplaxmi Patil, and Aamod Kale for their invaluable input. #genai #aiml #llm #langchain #insurance #costoptimization #grafana #bigquery #openai

36

1 Comment

🌟 Detecting Skin Cancer with CNNs: My Journey Towards Building a Robust Model 🌟 Skin cancer is one of the most common forms of cancer, and early detection can save lives. Inspired by this, I built a melanoma detection model using CNNs, leveraging the "Skin Cancer ISIC: The International Skin Imaging Collaboration" dataset. The dataset comprises 2,239 training images and 118 test images across 9 classes of skin lesions: Actinic Keratosis Basal Cell Carcinoma Dermatofibroma Melanoma Nevus Pigmented Benign Keratosis Seborrheic Keratosis Squamous Cell Carcinoma Vascular Lesion Here’s a summary of my progress and findings 👇 🔍 Model 1: Initial Approach Training Accuracy: 76.23% Validation Accuracy: Peaked at 53.47% Key Issue: Overfitting Validation accuracy stagnated, and validation loss increased while training metrics improved. Takeaway: A clear need for better generalization strategies. 🔧 Enhancements 1️⃣ Data Augmentation: Added augmentation techniques like rotation, zooming, and flipping to make the model more robust Result: Reduced overfitting but still observed class imbalance effects. 2️⃣ Class Rebalancing: Addressed class imbalance using Augmentor to ensure each class had equal representation: Generated 500 samples per class, significantly improving model balance. Improved generalization across all classes. 🚀 Final Model: Key Observations Architecture: 3 convolutional layers with ReLU activation Dropout layers (50% and 25%) to mitigate overfitting Adam optimizer and sparse categorical cross-entropy for multi-class classification Results: Training Accuracy: 89.54% Validation Accuracy: 86.93% Validation Loss: Stabilized at ~0.51 Overfitting: Minimal, with training and validation metrics closely aligned. 💡 Key Takeaways Data augmentation and class balancing are crucial for handling small, imbalanced datasets. Early stopping and dropout layers effectively prevent overfitting. Continuous monitoring of training/validation trends helps identify bottlenecks early. This project has been an enriching experience in using CNNs for healthcare applications. With more optimizations, this model could potentially assist dermatologists in identifying skin cancer early, enabling timely interventions. Project link: https://lnkd.in/gPt-u6p4 #MachineLearning #DeepLearning #HealthcareAI #SkinCancerDetection #AIForGood

20

2 Comments

I've recently been replicating some of our released work on LLM with MLX. I enjoy MLX for a variety of reasons. Minimalist is one of them. MLX's API is a subset of numpy and much less than PyTorch's. There was a moment when I thought the API was insufficient to support beam search and tree attention. However, as I moved along, I realized that this simple API is just as powerful as the ones on stage. Rethinking how to rewrite our PyTorch programs in MLX often resulted in simpler, more readable code. For example, the following two lines generate attention bias in PyTorch (https://lnkd.in/gszdymfE):   attn_bias = torch.zeros(causal_mask.shape, dtype=dtype, device=causal_mask.device)   return attn_bias.masked_fill_(causal_mask.logical_not(), LOG_0) MLX lacks masked_fill_. So I rewrote the above as a single line:   return mx.logical_not(causal_mask) * LOG_0 Yes. It takes time to rethink and rewrite code. However, if this work helps to keep MLX's API simple, it eases the learning process for others. I appreciate a small set of primitives. We would benefit from higher layers of APIs built on top of it, rather than a large flat set of primitives.

82

2 Comments

Intuit's recent decision to lay off developers, citing the need for more advanced AI skills, comes off as a disingenuous manoeuvre that sidesteps the truth. It's hard to buy into the narrative that these layoffs are truly about acquiring employees with better AI proficiency. In reality, it seems like a convenient cover for a cost-cutting strategy that involves offshoring jobs to cheaper markets. Think about it: the idea that current developers lack sufficient experience with AI tools like ChatGPT, Copilot, or other generative technologies is misleading. These tools have been part of the industry landscape for the past few years, and developers everywhere have been getting up to speed with them. The notion that there's a significant skill gap among developers when it comes to these tools is simply not accurate. The tools themselves, while helpful in certain contexts, are far from the magic solutions they are often hyped up to be. At best, they serve as fancy autocompletes—useful for streamlining some tasks but incapable of replacing the nuanced, expert work required for sophisticated software development. The real issue here is that Intuit's move undermines the trust and morale of its employees. Firing seasoned developers under the pretext of needing more AI-savvy replacements is not just misleading, it's a slap in the face to those who have dedicated their skills and efforts to the company. It creates an atmosphere of insecurity and distrust, where employees are left wondering if their hard work and loyalty mean anything in the face of corporate cost-cutting. Moreover, this strategy reflects a short-sighted understanding of what AI can achieve. While AI tools can certainly assist with certain aspects of development, they are not yet sophisticated enough to replace human expertise in creating complex, production-ready software. Believing otherwise is a risky gamble that could compromise the quality and innovation of Intuit's products in the long run. If the true intention behind these layoffs is to offshore jobs to cheaper markets, then Intuit should be honest about it. Hiding behind the AI narrative is dishonest and disrespectful to the employees who are being let go. Offshoring comes with its challenges, such as communication barriers and potential declines in work quality, and it is not the silver bullet it might appear to be. In an era where ethical business practices and corporate social responsibility are increasingly valued, such a strategy can erode customer trust and loyalty. #intuit #layoff #jobs https://lnkd.in/dv3qD9kS

4

1 Comment

I’m thrilled 🚀 to share my latest blog post: "𝐂𝐡𝐚𝐭𝐭𝐲 𝐆𝐢𝐚𝐧𝐭𝐬: 𝐓𝐡𝐞 𝐖𝐡𝐢𝐦𝐬𝐢𝐜𝐚𝐥 𝐖𝐨𝐫𝐥𝐝 𝐨𝐟 𝐋𝐚𝐫𝐠𝐞 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬!" 🌟 In this fun and informative dive, I explore how Large Language Models (LLMs) are transforming the way we write, code, and communicate. From chatbots that never leave you on read to coding assistants that feel like superpowers, these AI marvels are changing the game! 🔍 𝐖𝐡𝐚𝐭 𝐭𝐨 𝐞𝐱𝐩𝐞𝐜𝐭: - A straightforward breakdown of what LLMs are - Tips on when and how to use them - Resources—both free and paid! - The pros and cons of embracing this technology Whether you’re a tech enthusiast, a writer, or just curious about AI, this blog has something for everyone! Check it out and let’s chat about your thoughts! 💬👇 #AI #LargeLanguageModels #TechTrends #Innovation #ContentCreation #Coding #Chatbots #LLMs

5

📝 Exciting news! My latest blog post on Medium is now live! In this multi-series blog, I explore anomaly detection in any system. I share my experiences in designing and implementing an end-to-end solution for anomaly detection patterns in user access patterns. This series will cover a variety of topics, including understanding application events, optimizing data in Snowflake, developing machine learning models, and integrating with external applications. In the first blog post, I dive into everything about events and how they should be handled in a distributed system. Check out the full article here to learn more: https://lnkd.in/g8mjARdd I would love to hear your thoughts and insights in the comments section! Let's start a conversation. #events #eventing #outboxpattern #anomalydetection #machinelearning #dataoptimization #AWS #Snowflake

63

5 Comments

🎓 Is a Masters in USA in Computer Sciene/Data Science still worth it? There is a lot of speculation around this question, with clickbait news coming up with random facts and suggestions. The economy and job market has been tough in US and across the world in 2024 and there is uncertainties around visa and work permit. Lots of student are yet to land an internship or job offer with surmounting pressure of visa validity. I wanted to share some ground truths based on personal experience helping family members and friends currently navigating this journey. 📚 Education's True Value The worth of a Masters degree transcends job prospects - it's about deep subject matter expertise and intellectual growth. 🌟 Quality Matters US universities continue to offer world-class research facilities and educational excellence - choose institutions aligned with your specific interests and career goals. 💼 Career Perspective While Big Tech opportunities are attractive, don't make them your sole focus. Many tech giants now have robust operations in India. Evaluate if your goals can be achieved without a Masters in US. 🔍 Beyond Academics Your degree is just one part of the equation. Focus on: Building professional networks Participating in industry events Contributing to research Engaging in hackathons Exploring potential PhD paths 🚀 Pre-Masters Preparation Leverage online resources (MOOCs, YouTube) to build foundational knowledge before your program begins. This gives you a competitive edge for internships and job placements. 💰 Financial Planning Be realistic about the investment. While US opportunities are valuable, maintain a global perspective on career possibilities, including emerging markets. If you're passionate about your field and have done thorough research, don't let market uncertainties deter you. Success comes from strategic planning and staying adaptable. #HigherEducation #MastersDegree #STEM #InternationalEducation #CareerGrowth #USEducation #TechCareers #GlobalOpportunities #StudentLife #CareerAdvice #IndianStudentsAbroad #StudyAbroad #Education2024 #TechEducation #CareerDevelopment

14

I'm thrilled to be speaking about "Apache XTable (Incubating): Interoperability across Open Table Formats" at The Fifth Elephant 2024 annual conference tomorrow at BIC, Bangalore organized by Hasgeek In this talk you will learn about: 1. What are Table Formats? Why do we need Inter-Operability ? 2. A deep dive on XTable and problems that can be solved using XTable. 3. A demo of XTable in action. 4. The roadmap for the OSS project. You can check out and give a star to the OSS repository for the project here: https://lnkd.in/g53VWyU4 🎟️ Get your tickets and join us in-person: http://has.gy/sSCX P.S First 5 people to buy a ticket to the conference using the above link get a 10% discount! So go grab your ticket fast! #Hasgeek #AI #TheFifthElephant #MachineLearning #DataScience #BigData

85