Advertise on podcast: Data Forward: AI and Data Management News

Engineering YARN Optimization for High-Performance Hadoop Clusters In this episode, Ashwin Rajeeva (CTO & Co-founder, Acceldata) sits down with Pallav Kulshreshtha (Senior Director – Engineering, Acceldata) to unpack one of the most persistent challenges in Hadoop environments: resource wastage in YARN. Traditional YARN clusters allocate resources based on static, requested limits rather than real usage, leaving CPU and memory idle even when workloads don't need everything they've reserved. In this conversation, Pallav breaks down how a smart optimizer can close that gap using real-time telemetry, container fingerprinting, and adaptive overcommitment, all while keeping cluster stability intact. Whether you're managing a large-scale Hadoop deployment or exploring ways to squeeze more performance out of existing infrastructure, this episode offers a grounded, engineering-first look at turning YARN from a static allocator into a self-tuning, responsive resource manager. Learn more about Acceldata: engineering.acceldata.io. Subscribe for more engineering deep dives from the Acceldata team. #Hadoop #YARN #DataEngineering #ResourceOptimization #Acceldata #BigData #ClusterManagement #Podcast

In this episode, Ashwin Rajeeva sits down with Nikkitha Shanker, CEO of SuperBryn, to dig into what it actually takes to move voice agents from flashy demos to dependable production systems. They unpack why voice AI is fundamentally an orchestration problem, not a model problem, covering the real-time streaming architecture connecting telephony, ASR, LLM, and TTS, and why every millisecond in that pipeline matters. The conversation explores hard engineering problems like turn detection and barge-in, why time-to-first-audio (not raw inference speed) is the latency metric that defines user experience, and how geography and connection management quietly make or break a deployment. At the center of it all is SuperBryn's Ring System: a structured framework for classifying voice agent failures, conversational, task, policy, experience, and recovery, so teams can move from "the agent felt off" to "Ring 3 failed during interruption handling, here's who owns it." Ultimately, this is a conversation about trust: how reliability frameworks, continuous evaluation, and audio-native testing are what will separate voice agents that wow in a demo from those that survive the messy, unpredictable thousandth call, and become real infrastructure.

As enterprises race to deploy AI, many are discovering that success requires more than models and infrastructure. It demands trusted data, governed workflows, operational visibility, and measurable business outcomes. In this episode, Ramon Chen, Chief Product Officer at Acceldata, sits down with Lofan Leung, Director of Alliances at ServiceNow, to discuss how the partnership between ServiceNow and Acceldata helps organizations bridge the gap between data, AI, and business execution. The conversation explores ServiceNow's Workflow Data Fabric and AI strategy, including how enterprise data, metadata, workflows, and AI agents come together to create intelligent operations. Lofan also shares insights into ServiceNow's partner ecosystem and how organizations are leveraging the platform to accelerate digital transformation and AI adoption. The discussion then turns to the growing partnership between ServiceNow and Acceldata. Together, ServiceNow's Workflow Data Fabric and Acceldata's Data, Compute, and AI Observability capabilities provide organizations with the trust, visibility, governance, and operational intelligence needed to scale AI responsibly. From data quality and lineage to AI agent monitoring, governance, and auditability, the combined solution helps enterprises move from experimentation to production with confidence. Whether you're a CIO, CDO, data leader, or AI executive, this episode offers a practical look at how trusted workflows and trusted data can accelerate enterprise AI outcomes.

Recorded live at Autonomous 26 in San Francisco, Ashwin sits down with Rajesh Vijayaraghavan, Principal ML & AI Specialist at AWS, to unpack how enterprises and ISVs are moving from AI experimentation to real production value. AWS's Bedrock and Agent Core platform gives companies a modular stack — covering model access, memory, observability, and multi-agent orchestration — without managing underlying infrastructure. ISVs like Box, Asana, and Cox Automotive are embedding agentic workflows to keep users in their apps and automate cross-platform tasks, with some running hundreds of agents in production. After years of "leap of faith" AI investment, the last six months have shown measurable productivity gains — with AWS's own Mantle engine built by fewer than 10 engineers in weeks using AI coding tools.

How do you connect thousands of distributed data points across multi-cloud, hybrid, and edge environments without spending months on firewall tickets and VPN configurations? In this technical deep dive, Gowtham Saasivam - the Senior Staff Engineer at Acceldata, explains how the "Cloudbridge" architectural pattern is deconstucted —a novel approach to enterprise connectivity designed for planetary-scale distributed systems. Traditional networking (IP VPNs, manual TLS, and complex firewall rules) creates massive operational bottlenecks and security risks. We explore how to flip the script by moving from network-centric connectivity to identity-centric, automated secure channels. What we cover in this session: * The Death of the Inbound Port: Why traditional "inbound" firewall rules are a bottleneck for CI/CD and how Reverse Connectivity allows data planes to initiate outbound-only mTLS tunnels. * Automating the PKI Nightmare: How to implement a zero-touch certificate lifecycle that eliminates manual issuance and prevents outages caused by expired certs. * Zero-Trust in Motion: Moving beyond the initial handshake to implement continuous authorization and behavioral monitoring for every request. * Engineering Resilience: An inside look at adaptive backoff algorithms and dynamic service discovery for high-availability tunnels. Key Engineering Takeaways: 1. Reverse Connectivity: Use outbound-initiated channels to bypass complex ingress policies. 2. Identity-Driven Routing: Leverage automated mTLS certificates for both security and service discovery. 3. Operational Autonomy: Reduce the "coordination tax" between NetOps, SecOps, and App Dev teams. Who is this for? Distributed systems engineers, Cloud Architects, SecOps professionals, and anyone building global data platforms that need to reach into highly segmented or restricted environments. #CloudNetworking #ZeroTrust #DistributedSystems #DataEngineering #CyberSecurity #mTLS #DevOps #CloudBridge #SoftwareArchitecture #EnterpriseIT

Abstract:In this episode of Data Forward, Ramon Chen sits down with Guy Voster to explore why AI initiatives are failing—not due to a lack of data, but a lack of context. As enterprises race to adopt AI, the conversation is shifting toward context intelligence—understanding where data comes from, what it means, and whether it can be trusted. Ramon and Guy unpack how unified, entity-centric data combined with semantics and real-time delivery creates a reliable foundation for both humans and AI agents. They also dive into: Why AI without context leads to confident—but incorrect—decisionsThe role of master data management and data observability in building trustHow fragmented data environments create risk for AI-driven automationWhat it takes to move from raw data to actionable, context-rich intelligence This episode is a must-listen for data and AI leaders looking to bridge the gap between data, meaning, and action in the age of intelligent systems.

Abstract: On this Data Forward episode, Jason Russler of VAST Data and Ramon Chen of Acceldata discuss how enterprises can finally modernize their legacy Hadoop environments, cutting costs dramatically while positioning their data infrastructure for the AI era. Listeners will learn: How swapping HDFS for the VAST AI OS with S3A-compatible object storage eliminates 3× replication overhead and shrinks storage footprints by up to 8× — without sacrificing performance Why stable workloads like Spark ETL are moving back on-premises as cloud bills soar, and how private AI environments deliver both cost savings and data sovereigntyHow Acceldata's observability tooling surfaces hidden cluster inefficiencies, turning existing hardware into found money before any new infrastructure spend

What if Kubernetes releases moved like messages on a bus — immutable, traceable, and fully observable? In this episode, Ashwin Rajeeva ft. Neeraj Nayan, Senior Kubernetes experts from Acceldata, unpack the architecture behind Flow Controller — a system designed to eliminate environment drift and bring identity, immutability, and auditability to Kubernetes releases at scale. If you’ve ever heard “it worked in dev” and knew the code wasn’t the problem, this conversation will resonate. The real issue is often configuration drift, manual production fixes, broken promotion chains, and reconstructed intent. Adding more pipeline checks doesn’t solve that. Identity does. Ashwin and Neeraj break down: * Why traditional Kubernetes promotion workflows fail at scale* The concept of a Release Bus and immutable release descriptors* How UUID-based release identity removes guesswork from promotions* Designing a GitOps-native system that works with ArgoCD, not against it* Template-driven release generation to prevent copy-paste drift* Real-time promotion tracking for incident-time clarity* How to definitively answer: “Is production running exactly what we tested?” Flow Controller doesn’t replace Git or ArgoCD. Git remains the source of truth. ArgoCD remains the reconciler. What changes is the orchestration layer — adding traceability, structured promotions, and verifiable release identity across environments. This episode is for platform engineers, DevOps leaders, SREs, and Kubernetes practitioners building at scale — especially those navigating GitOps workflows, release governance, and production reliability. Because velocity without identity is just faster confusion. And when releases carry identity end-to-end, Kubernetes finally becomes predictable.

In this first episode of the Acceldata Engineering Podcast, Acceldata founders Ashwin Rajeeva (CTO) and Raghu Mitra (Co-founder & VP Engineering) sit down to unpack how AI is transforming Agentic Data Management (ADM) and modern data engineering. They share real, first-hand insights on:1. How AI fits into today’s data engineering and analytics stacks2. The role of ADM in building reliable, scalable data platforms3. Where abstraction layers help—and where they hurt4. What founders think differently about when building data platforms with AI at the coreThe future of AI-driven data observability, reliability, and operations This conversation goes beyond buzzwords, diving into practical engineering decisions, architectural trade-offs, and real-world lessons from building Acceldata. Whether you’re a data engineer, platform architect, analytics leader, or founder, this episode offers a grounded perspective on AI, ADM, and the evolving data ecosystem—straight from the brains behind the platform. Watch now to hear how Acceldata’s founders think about building data platforms in the age of AI.

🚀 Data Forward: Building Reliable Data Systems in the Age of AI In this episode, Joe Murphy, Principal Architect at Acceldata sits down with Jay Sivasamy, Principal, Enterprise Data & AI at Workday to explore how enterprises can build trustworthy, reliable data systems in today’s AI-driven world. They dive into: 🔹 Evolving Frameworks: From ETL scripts to AI-ready pipelines 🔹 Trust & Governance: Making reliable data a business-critical asset 🔹 AI Integration: Using AI agents to maintain and improve system reliability 🎧 Tune in and learn how to future-proof your data systems: https://linktr.ee/dataforward #DataForward #DataReliability #AI #EnterpriseData #DataGovernance #MachineLearning #DataEngineering

In this episode of Data Forward, host Amy Kyleen Lute talks with Mahesh Kumar, CMO at Acceldata, about the findings of the new AI Readiness Benchmark Report - based on responses from 150+ data leaders at large enterprises. Mahesh shares why AI readiness is critical, how organizations can measure their preparedness, and the risks when data foundations lag. On this podcast you will learn: -The key metrics to use to measure AI Readiness -The contextual indicators that can help improve your key metrics -How using AI to improve data operations can give organizations a competitive edge

In this episode of Data Forward, Ramon Chen, CPO of Acceldata talks with Matthew Cox, CDO at Fusable, about turning data into a true enterprise asset. They cover how Fusable unifies siloed data, ensures trust and quality with observability and lineage, and prepares for the next wave of AI innovation. Key Takeaways -Unify Data to Unlock Insights – How Fusable delivers value by creating a single, trusted data core. -Trust & Quality First – Why lineage, observability, and governance are essential for reliable data. -AI Readiness Starts with Data – Why high-quality, well-structured data is key for Gen AI and Agentic AI.

In this special episode of Data Forward, Vivek Singh sits down with Chandrakant “Chandra” Sharma — Acceldata’s Global Field CTO and VP of Field Engineering — for a candid, practical conversation on how the world’s largest enterprises are modernizing their data platforms. Chandra shares proven, real-world strategies for: Simplifying complex, fragmented data stacks Avoiding vendor lock-in while controlling cloud costs Safely migrating off aging Hadoop platforms Building AI-ready, reliable platforms that deliver business value Whether you’re a data leader, architect, or engineer, you’ll walk away with actionable lessons you can start applying immediately to improve agility, reduce costs, and future-proof your data platform.

In this episode, Ramon Chen sits down with industry veteran Dan Power to explore the evolution of data governance, quality, and management in the age of AI. From foundational lessons at Dun & Bradstreet to regulatory pressures in financial services, Dan shares what it takes to build trusted, scalable data practices—and where AI fits into the picture. Key Takeaways:- Data Products Need Discipline: Creating reliable data products requires manufacturing-like thinkingclear inputs, measurable outputs, and trust scores to drive business value.- AI Is Not a Silver Bullet—But a Power Tool: GenAI and agentic systems can automate manual data work, but only if foundational data observability is in place.- Regulations Demand Operational Readiness: Regulatory compliance increasingly depends on fast, accurate lineage, quality metrics, and the ability to prove control—not just documentation.