Pluang is Mastering with Redis
Stay tuned with Aditya Jha, CTO at Pluang and Ankit Agarwal, VP Engineering at Pluang and Jatin Gupta, Regional Director at Redis APAC in the episode of our series #CustomerChronicles.
Note: This video is a part of ETCIO's Brand Connect Initiative.
Advertisement
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
a day ago
- Time of India
AI for the masses: Inside Meesho's bold leap with BharatMLStack
Artificial Intelligence is no longer a back-end tool—it sits at the very heart of digital commerce, shaping how technology communicates, recommends, and influences consumer decisions. Underlining this, a report titled EY Consumer Index Survey sheds light on India's readiness for this shift: 82% of Indian consumers believe AI will significantly improve their shopping experience, while 30% say they understand AI, compared to just 17% globally. Surprisingly, despite being an emerging tech market, Indian consumers are outpacing their global counterparts in AI awareness and trust—especially in areas like personalized recommendations, where 48% of Indians trust AI to tailor deals and offers, double the global average. To explore how AI is being scaled, democratized, and personalized for 'Bharat,' ETCIO interacted with Debdoot Mukherjee, Head of AI & Demand Engineering at Meesho , one of India's fastest-growing e-commerce platforms . Edited Excerpts: Meesho recently open-sourced the BharatMLStack on GitHub, becoming one of the first Indian e-commerce platforms to do so. What strategic vision underpins this move, and how do you see it shaping India's AI ecosystem, especially in Tier 2+ markets? The decision to build and eventually open-source BharatMLStack came from our need to develop a robust internal ML platform to support hundreds of diverse use cases at scale. We needed something cost-efficient, reliable, scalable, and extensible—and over time, our in-house stack evolved to serve 66.90 trillion of feature retrievals and real-time predictions. That's when we decided to open-source BharatMLStack—to help startups and larger companies leapfrog their AI journeys. It's especially useful for those building for Bharat, where scale and cost efficiency are critical. On that note, how is Meesho leveraging AI to power smarter, more intuitive shopping experiences for mobile-first, value-conscious users in Bharat? Can you share some use cases that show direct impact? Absolutely. For users from Tier 2+ towns, there are barriers to online shopping—language, ambiguity in search, and lack of digital familiarity. AI helps break these barriers. One major area is personalization—we use deep learning models across surfaces like homepage feeds and search to serve relevant products based on user behavior and intent. On the seller side, our AI helps ease cataloging through image recognition and enrichment, automates pricing insights, and improves product visibility—especially critical for first-time or non-tech-savvy sellers with razor-thin margins. Are there any unique challenges or breakthroughs your team has encountered while scaling these AI systems? Yes, several. One major challenge is interpreting long-tail, ambiguous search queries, often typed in Indian languages using English scripts or even voice. We've used LLMs to decode intent and improve search relevance. Another is balancing relevance with fairness—ensuring that newer sellers also get visibility despite algorithmic biases like the 'rich get richer' effect. Our ranking systems are designed to support platform fairness, long-term seller success, and ultimately, greater consumer choice and price competitiveness. With AI becoming more influential in consumer apps, how do you approach responsible AI, bias mitigation, and compliance, especially in light of India's evolving data privacy regulations? Data privacy is something we take very seriously—we're fully aligned with current and upcoming regulations. But beyond compliance, responsible AI is embedded into our philosophy. We constantly monitor for algorithmic bias, especially in recommendation systems, and work to counteract unintended patterns that could marginalize certain sellers. This isn't just good ethics—it's good business. A healthy, competitive seller base offers better prices and more variety, benefiting the platform and our users. Looking ahead, what are some areas of AI innovation you and your team are most excited about—LLMs, multimodal models, or deeper integrations in logistics, seller support, and beyond? There's so much happening, it's hard to choose! But yes, GenAI and LLMs are becoming central. We've implemented GenAI-powered customer support chatbots and are expanding that across user categories. In logistics, we've built an AI-first platform called Valmo, where everything from delivery planning to fraud detection is AI-powered. For instance, we created a Geo-India LLM to decode vague, rural delivery addresses—something no other off-the-shelf tool could do. It's a great example of multimodal intelligence tailored for India.
Time of India
5 days ago
- Time of India
Pluang is Mastering with Redis
Real-time trading isn't just about speed. It's about precision, resilience and innovation at scale. Stay tuned with Aditya Jha, CTO at Pluang and Ankit Agarwal, VP Engineering at Pluang and Jatin Gupta, Regional Director at Redis APAC in the episode of our series #CustomerChronicles. Note: This video is a part of ETCIO's Brand Connect Initiative. Advertisement
Time of India
26-06-2025
- Time of India
Data democracy: Why IT must lose control for businesses to win
As enterprises race to become data-first, CIOs are shifting from IT leaders to business enablers—breaking silos, simplifying access, and embedding data-driven culture across the ETCIO Annual Conclave 2025, the provocative trilogue chat 'Data Democracy: Why IT Must Lose Control for Businesses to Win' set the stage for a brutally honest conversation on reimagining the data-to-decision pipeline in modern enterprises. Moderated by Sneha Jha, Editor at ETCIO, the session featured compelling perspectives from Vishal Gupta (Head of Business Transformation & Digital, Godrej Consumer Products), Pavan Goyal (CIO, Mphasis), and Pramod Mundra (President & CIO, Havells). Together, they challenged the traditional power equations between IT and business and offered an insider's lens into how progressive enterprises are redistributing the power of data. Democratizing without diluting: The quality imperative Vishal Gupta framed the discussion by stating the uncomfortable truth: 'Democratization of bad data is worse than not democratizing at all.' For data democracy to deliver value, enterprises must first ensure a single, well-governed source of truth. At Godrej Consumer Products, this means centralizing data platforms, ensuring transformation integrity at the source, and putting data accountability on application owners. 'Cutting access to alternate data sources forces adoption and surfaces quality issues faster,' he emphasized. Pramod Mundra echoed the sentiment, underscoring the importance of allowing teams to fail fast—if they can recover faster. Drawing on Amazon's 'one-door vs. two-door decisions' framework, he advocated calibrating the freedom to experiment based on reversibility. 'Let people own their analytics. Just ensure everyone refers to the same version of truth,' he said. When experience meets data—and friction ensues Democratization often leads to political turf wars when data challenges gut instincts. In legacy-heavy industries like FMCG and manufacturing, this tension is palpable. 'Traditional managers trust their experience more than the data,' Pramod admitted. The solution? Make executive reviews dependent on standard dashboards and ensure data access is frictionless. 'Friction becomes an excuse to resist change,' he warned. With top-down sponsorship and seamless access, even the most change-averse departments begin aligning with data-led decision-making. Breaking silos, building platforms, encouraging playgrounds Pavan Goyal took the conversation further by exploring both the pros and cons of silos. 'Not all silos are bad,' he remarked, drawing a distinction between counterproductive knowledge silos and value-adding specialty units. To dismantle the former, Mphasis deployed an AI-enabled knowledge management framework—gamifying contributions and rewarding knowledge sharing. 'Our content consumption jumped 10x in six months,' he revealed. Empowering business teams through AI-powered data platforms and self-service tools emerged as a clear strategy across panelists. Havells, for example, is creating a language model using Databricks to enable enterprise-wide conversational access to data. 'This will drastically reduce friction,' said Pramod. Meanwhile, Godrej is enabling teams to move from univariate to multivariate analyses to improve the precision of business insights. Perhaps the most powerful metaphor of the evening came from Vishal Gupta: 'The CIO must evolve from being a gatekeeper to becoming a shopkeeper of insights.' Like a smart retailer, the CIO must ensure that the right data is on the right shelves at the right time—ready for consumption, without barriers. He advocated moving from a 'Kirana shop' model—where the IT team curates data—to a 'marketplace model'—where business users co-create data experiences. 'It's not just about access. It's about creating a pleasurable, productive experience with data,' he said. Guardrails, governance, and cultural realignment When asked about reconciling business agility with governance, the panelists were unanimous: data governance is non-negotiable. Pramod emphasized that it's not IT's mandate alone but an enterprise-wide commitment. 'Overriding privacy and security is not an option,' he said. The panel also highlighted that governance needs to be embedded in organizational culture, not layered on as an afterthought. From access to impact: Stories that sell the dream The panelists shared powerful anecdotes of how data access at the mid-level has led to high-impact decisions. Vishal recalled how Godrej's marketing team used granular data to correlate commodity price shifts with competitor pricing and internal profitability—turning data into a strategic lever. However, he warned that without the right analytical training, flawed assumptions—like simplistic univariate analyses—can lead to misleading insights. Hence, the need for guardrails alongside empowerment. Monetizing data: From hype to meaningful impact The session concluded with a reality check on data monetization. Rather than external sales of data, each panelist focused on internal value creation—be it enabling hyper-personalized customer experiences, accelerating project kick-offs, or enhancing distributor efficiency. 'We're not selling data. We're selling better decisions,' Vishal noted, as the audience nodded in agreement.
