Day 1 of DataHack Summit 2025 set the inspiration for India’s most futuristic occasion. In case you missed it, you possibly can learn all concerning the DHS Day 1 happenings right here. On August 21, Day 2 of DHS 2025 constructed upon it with daring concepts, cutting-edge demonstrations, and a renewed imaginative and prescient for synthetic intelligence. With classes spanning belief, inclusivity, creativity, and robotics, the second day of the summit reminded everybody why DHS is the beating coronary heart of India’s AI revolution.
DHS 2025, Day 2: AI Specialists Assemble
The day opened with a keynote speech by Joshua Starmer, Founder & CEO of StatQuest. Simply as at all times, Starmer captured everybody’s consideration together with his ukulele efficiency and riveting session on Quantifying Our Confidence in Neural Networks and AI. By displaying how confidence might be measured, from easy neural networks to at this time’s advanced LLMs, he offered a framework for belief in methods that too usually produce outputs with out accountability. His strategy reminded the viewers that AI’s future is extra about reliability than simply energy.
Carrying that momentum ahead, Manish Gupta, Senior Director at Google DeepMind, explored the huge potential and urgent challenges of Inclusive AI. His discuss spanned the promise of basis fashions like Gemini to their real-world utility within the World South, from personalised studying to agricultural transformations utilizing satellite tv for pc imagery. His message was clear: AI should not solely be cutting-edge, it should be inclusive, honest, and constructed for billions.
Past Knowledge to Discovery
The highlight then turned to Abhishek Divekar, Senior Utilized Scientist at Amazon. He addressed the viewers with The Promise and Pitfalls of Artificial Knowledge Technology. Whereas artificial knowledge opens doorways to privacy-conscious, various coaching units, he cautioned towards careless scale that amplifies hallucinations and biases. His insights grounded the viewers within the delicate steadiness between innovation and accountability.
One more highly effective session got here from Alessandro Romano, Senior Knowledge Scientist at Kuehne+Zagel. He bridged the hole from LLMs to Agentic AI, displaying how multi-agent methods can resolve dynamic, real-world issues. Utilizing frameworks like CrewAI, he demonstrated how collaborative, goal-driven brokers mark a step-change from static immediate engineering to clever orchestration.
The position of the information scientist is evolving quickly, and therefore, a change is essential. Anand S from Straive make clear the identical, delivering a thought-provoking Hack Session titled “RIP, Knowledge Scientists.” This wasn’t a farewell however a wake-up name, a daring take a look at how automation and AI are remodeling conventional knowledge roles. Automating repetitive knowledge wrangling and inventive, strategic problem-solving are the brand new norm. His message was clear: knowledge scientists who embrace this shift will thrive. For anybody navigating the tides of the data-driven world, this was a blueprint for the long run.
Vibe Coding Showdown
The thrill peaked within the Vibe Coding Showdown. Anuvrat Parashar, Anand S, and Ravi RS Nadimpalli went head-to-head in constructing purposes with AI assistants. Of their electrifying Hack Panel on “Vibe Coding Showdown: Constructing Purposes with AI Assistants,” the trio displayed collaboration at its most interesting – a masterclass in mixing human creativity with AI precision. For builders and tech fans within the viewers, it supplied a glimpse into the way forward for coding: a dynamic partnership that transforms how purposes are imagined and constructed.
Constructing Smarter, Quicker, Safer
Sensible purposes with Pankaj Agarwal, Senior Software program Engineer (Machine Studying) at Uber, dominated the afternoon classes. He showcased how RAG-powered search question optimization can rework obscure, noisy meals supply queries into exact, intent-aware outcomes. In the meantime, Dr. Kiran R, VP of Engineering at Oracle, offered a deep dive into Evaluating GenAI Fashions. Dr Kiran framed the significance of sturdy metrics, ground-truth datasets, and human-in-the-loop practices in enterprise and healthcare contexts.
Including to the momentum, Nitin Agarwal, Principal Knowledge Scientist at Toast, and Rutvik Acharya, Principal Knowledge Scientist at Atlassian, defined how the Mannequin Context Protocol (MCP) is streamlining software integration for brokers. Their reside demo showcased how MCP can scale back overhead, standardize workflows, and unlock extra scalable automation.
Different classes pushed boundaries additional. Anand S, LLM Psychologist with Straive, sparked debate together with his provocatively titled discuss, RIP, Knowledge Scientists, displaying how LLMs can now carry out a lot of the normal knowledge scientist’s workflow, and what roles stay important for people. Krishna Kumar Tiwari, Co-Founder & CTO at Whitler AI, in the meantime, demonstrated how GenAI brokers are revolutionizing efficiency advertising at scale, crafting thousands and thousands of personalised creatives in actual time.
Pushing the Inventive Edge
From Web3 to sustainability, Day 2 proved AI’s versatility. Knowledge scientist Rohan Rao confirmed how GenAI is reimagining the Web3 ecosystem, powering NFTs, sensible contracts, and crypto safety. Ruchi Awasthi, Machine Studying Engineer – CTO Workplace at Pure Storage, tackled the urgent query of sustainability in her discuss on LLM Compression. Ruchi broke down strategies like pruning, quantization, and data distillation to make AI extra environment friendly and eco-friendly.
Hitesh Nayak, Senior Director of Knowledge Sciences at Determination Foundry, explored how MCP-driven brokers can rework advertising. Harshad Khadilkar, Lead Knowledge Scientist at Tata Group, took a inventive flip together with his session ‘LLMs Are Boring. How Can We Make Them Extra Attention-grabbing?’ His name for directed randomness and novelty reminded the viewers that discovery, not simply summarization, ought to be the north star of GenAI.
The narrative deepened with Pavak Biswal and Abhilash Kulkarni of Dentsu World Providers, who unveiled Saving Ananya, a model AI playbook for buyer retention. Via a demo-driven story of predictive capabilities, threat scoring, and agentic triage, they confirmed how manufacturers can rework ache factors into loyalty. Complementing this, Karan Sindwani, Senior Utilized Scientist at AWS, revealed how GenAI is powering production-ready options throughout industries. From customer support to cricket analytics, GenAI is proving to be the artwork of potentialities.
From Robotics to Full-Stack Brokers
One of many day’s most inspiring moments got here from Logesh Kumar Umapathi of BLACKBOX.AI, who offered From Language to Robotics. His reside demo with robotic arms illustrated how LLMs and reinforcement studying are converging to make AI actually embodied. The viewers witnessed firsthand how idea interprets into tangible, bodily affect.
Equally forward-looking had been talks by Daksh Varshneya, Senior Product Supervisor at Rasa. Daksh launched Course of Calling as a superior paradigm for conversational brokers. Saurav Agarwal, Options Structure and Engineering at NVIDIA, showcased the Full-Stack Agentic AI toolkit to speed up, consider, and scale agent methods. Miguel Otero Pedrido, ML Engineer and Founder at The Neural Maze, then took the viewers past PoCs together with his speech. He demonstrated tips on how to productionize agentic methods with strong monitoring and LLMOps practices.
Closing out the technical monitor, Ayush Thakur, Machine Studying Engineer at Weights & Biases, unveiled The Lacking Piece of AI Apps: Analysis. His hands-on methods for constructing evaluators, aligning them with human preferences, and scaling bias evaluation proved that analysis is a characteristic and never an afterthought.
DHS 2025: Day 2 Award Winners
Amid business leaders, the night was additionally marked by celebration. AV Luminary Awards honoured the High 7 GenAI Scientists for his or her groundbreaking contributions:
- Abhishek Divekar
- Amit Sharma
- Karan Sindwani
- Logesh Kumar Umapathi
- Mayank Baranwal
- Sunny Manchanda
- Tanika Gupta
Every recipient embodied the innovation and dedication driving India’s AI motion ahead. The corridor stuffed with applause for the collective brilliance and spirit of this group.
DHS 2025: A Summit of Concepts, Power, and Influence
Day 2 of DataHack Summit, or DHS 2025, was a peek at how the AI group is shaping the long run. From confidence estimation in neural networks to embodied robotics, from sustainability in LLMs to agent orchestration at scale, the classes illustrated a outstanding spectrum of progress.
With two days nonetheless to go, the summit’s journey has solely simply begun. Keep tuned to this house as India’s most futuristic occasion additional unfolds.
Login to proceed studying and luxuriate in expert-curated content material.
