Day 3

This talk explores how Generative AI and advanced machine learning techniques can revolutionize software engineering practices to build a learning organization. By leveraging structured and unstructured datasets, Generative AI enables predictive modelling, optimization, and continuous improvement.

The rise of AI is reshaping the programming landscape, presenting both challenges and opportunities for developers. In a country like India, where the middle class has been largely built on the IT services industry, this shift carries significant implications.
In this session, Tanay Pratap will explore how AI will impact the role of programmers and the broader tech ecosystem. He’ll share insights on:

How to be a better programmer with AI: Leveraging AI tools to automate repetitive tasks, enhance productivity, and focus on creative problem-solving.
How to build AI-powered apps: A practical guide to using AI services to create next-generation applications that are intelligent, efficient, and user-centric.
The session will also discuss how programmers and tech professionals in India can adapt to this new paradigm, ensuring they remain relevant and competitive in an AI-driven world. This talk is for developers, entrepreneurs, and tech enthusiasts who want to embrace the future of programming with confidence and purpose.

Join us as we delve into how Wakefit, India’s leading sleep and home solutions company, has transformed its business using cutting-edge technologies like Artificial Intelligence (AI), Machine Learning (ML), and Generative AI. This talk will showcase real world learnings on harnessing the power of these innovations to revolutionize both product offerings and customer experiences.

We will explore how we leverage AI/ML algorithms to optimize demand planning and streamline fulfillment processes. By analyzing customer purchase data, AI algorithms predict product demand, allowing the company to make informed decisions on inventory levels. This strategic approach avoids both overstocking or stockout and helps to ultimately drive business efficiency.

In addition, we will showcase real-world learnings from implementing GenAI to elevate the customer experience while reducing operational costs and increasing contact center efficiency. With this technology, Wakefit is redefining its customer interactions, resulting in a more personalized and seamless experience for its customers.

This research introduces a unified platform for fine-tuning multiple open-source AI models on small datasets, optimizing resource usage and ensuring data privacy. The approach employs Parameter-Efficient Fine-Tuning (PEFT) with QLoRA, dynamic hyperparameter tuning, and stratified random sampling for efficient performance. It automates data preprocessing, prompt adjustment, and compliance with responsible AI practices. By evaluating models on metrics like accuracy, F1 score, and perplexity, the platform provides actionable insights, reducing costs and time while enabling businesses to select the best model for their needs.

World has moved from drag and drop timeline based editors. Text to video is a paradigm which makes video creation and distribution a verbal skill. This takes this confined ability to the hands of every person who can think creatively and has a story to tell. But chatting with a video creation & editing agent has its own set of challenges, especially for the user asking for high content control as well as ease of creation. Existence of 1000+ languages poses another dimension in these challenges. BingeClip AI is configuring a set of technologies which include Agentic RAGs, Automated document workflows, Proprietary video search models, autonomous agents for editing operations and model compression. The combination creates a seamless application layer which enables user interaction, hence video creation on the move, combining authenticity of UGC with creativity & Automation of AI.

Integrating LLMs into workflows and interfaces can really enhance customer experience. However, achieving impactful outcomes is not an overnight transformation. Drawing from examples of our adoption process, this session will walk through the iterative journey of enhancing customer-facing interfaces, support agent tools & internal processes using LLMs.
The emphasis will be on:
Starting small and scaling thoughtfully.
Learning from each phase of implementation.
Continuously refining based on user feedback and performance metrics.

As AI becomes increasingly crucial to enterprise success, we see companies moving deeper into a technological frontier that promises both opportunity and risks. The advent of generative AI has driven a surge in AI use cases throughout organizations. So, how are companies preparing—or not preparing—to keep pace with this rapidly evolving technology and avoid unintended consequences, both known and unknown?

Join the session to get answers on below:

1. Who are the executives representing organizations with larger AI governance budgets?
2. Do organizational investments in AI governance reflect a sense of urgency?
3. How many new AI models are in the enterprise forecast?
4. What kinds of AI tools and frameworks are organizations using?
5. Are respondents confident that their AI capabilities can meet the challenges ahead?
6. Which AI risks and regulations worry enterprises the most? 7. Who owns organizations’ AI governance?

This research presents a transformer-based approach for Next Basket Recommendation (NBR) in grocery retail, focusing on predicting new, previously unpurchased products. Using a modified T5 architecture, the model captures item-to-item correlations across sequential shopping baskets, framing NBR as a sequence-to-sequence task. Extensive evaluations on the Dunnhumby dataset show that the model outperforms traditional collaborative filtering methods, achieving high precision and ranking effectiveness. This approach demonstrates the potential of transformers for personalized and novel product recommendations.

This research work presents Uniform Discretized Integrated Gradients (UDIG), a novel attribution method for explaining large language models by employing a nonlinear interpolation path optimized for discrete embedding spaces. Evaluated on sentiment classification (SST2, IMDb, Rotten Tomatoes) and question answering (BERT fine-tuned on SQuAD), UDIG demonstrates superior performance across log odds, comprehensiveness, and sufficiency metrics, surpassing traditional methods in predictive language model interpretability.