Why Enterprises Struggle to Operationalize AI: Last-Mile Failures, Data Quality, RAG Brittleness & ROI RealityIn this episode of “Techie Talks with Kuldeep,” host Kuldeep Singh speaks with Farshid Sabet, CEO and co-founder of Corvic AI, about why enterprises still struggle to operationalize AI despite major investment and excitement. Farshid shares his journey from hardware and edge AI to software and AI, then explains the “last mile” problem: prototypes impress, but production breaks in mission-critical settings where errors and hallucinations are unacceptable. They discuss brittle RAG and data pipelines at scale, rising costs, the need for data cleaning and consistent reporting, and how incomplete or fragmented data leads to hallucinations and “garbage in, garbage out.” The conversation highlights governance, security, privacy, orchestration across tools and modalities, and why AI ROI lags while the industry remains in a learning and build phase.Time Stamps:0:00 - Precaps1:30 - Introduction & Guest Welcome1:54 - Farshid's Journey: SanDisk to Corvic AI3:23 - The 3 Waves of AI (Audio → Vision → LLMs)4:28 - Why Enterprises Still Struggle to Operationalize AI6:45 - Mission-Critical AI: The Bar Is Very High8:13 - The Last Mile Problem & AI Hallucinations9:21 - Vibe Coding & Brittle Data Pipelines at Scale11:35 - Smarter Models Won't Fix Bad Data14:51 - From Dashboards to Live AI Analytics18:14 - Security, Privacy & AI Governance21:12 - Why AI ROI Is Still Lagging24:22 - The AI Hype Trap & the "Push a Button" Myth27:53 - The Hidden Cost Problem Nobody Talks About29:02 - Underrated Tech: Orchestration, Graph & Multi-Modal Data35:48 - Enterprise AI ≠ Consumer AI36:29 - Final Thoughts & Wrap-UpConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Farshid Sabet and Corvic.ai Linkedin : https://www.linkedin.com/in/farshid-sabet-298526/Corvic.ai : https://www.linkedin.com/company/corvic-ai/

AI Voice Agents in India: Girikon.ai’s Vernacular Voice Tech, CRM Automation & Future of Customer SupportIn this episode of “Techie Talks with Kuldeep,” host Kuldeep Singh interviews Alok, DC alumni and founder of Girikon.ai, about AI voice agents integrated with telephony (CTI) and CRMs like Salesforce/HubSpot to reduce IVR wait times, automate workflows, summarize calls, guide human agents with prompts, and transfer calls with full context. Alok explains Girikon’s shift from telephony to enterprise AI voice solutions, highlights support and sales use cases, and describes adoption in real estate, healthcare, and BFSI (e.g., EMI/payment reminders). He emphasizes that success depends on solving real business processes and having data in place, not just adopting tools. Girikon is building “homegrown” Indian-language voices (10+ languages) to address gaps in vernacular speech and improve accessibility, including rural use cases.Time Stamps:00:00 Precaps01:00 Show Intro and Guest01:18 Alok’s Founder Journey01:56 Giricon.ai and CTI Roots04:59 Enterprise Call Pain Points07:26 AI Agent in Action08:54 Seamless Human Handoff10:59 Indian Language Challenge13:26 Voice vs Chatbots17:45 Voice First Future22:10 Industry Use Cases24:26 ROI and Fast Deployment27:43 Three Week Deployments28:11 Why AI Projects Fail30:27 Scaling With Customer Data31:27 Real Time Agent Coaching33:00 Next Five Years Outlook35:25 Cost Of Voice Tech38:46 Homegrown Indian Voices40:37 Rapid Fire Takeaways42:53 Hiring And Retention46:48 Discovery To Adoption49:42 Global Delivery Model51:10 Closing ThoughtsConnect with Techie Talks with Kuldeep:YouTube Channel:  @techietalkswithkuldeep  WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Alok and GirikonLinkedin : https://www.linkedin.com/in/alokanibha/Girikon: https://www.linkedin.com/company/girikon/#aivoice #voiceai #aiagents #conversationalai #contactcenterai #callcenterautomation #crm #salesforce #workflowautomation #indianlanguages

Will AI replace data engineers? 🔥Why Data Engineering Still Matters in the Age of GenAI, AutoML & AI Agents. Host Kuldeep Singh interviews senior data engineer Meghana Ashok Ganatra on why data engineering remains critical despite hype around LLMs, GenAI, AutoML, and AI agents. Meghana explains that AI outputs depend on reliable, consistent data pipelines, and bad data can cause costly and harmful decisions, illustrated with a healthcare misdiagnosis example. They discuss how the data engineer role has expanded beyond ETL to working with data scientists on deployment, governance, and reliability, noting that “AI is the new electricity” still needs infrastructure. Meghana shares an EV telemetry case study migrating from Azure Synapse to Databricks to address terabytes of per-minute charger data, reducing costs and improving scalability and performance, using Unity Catalog for governance, handling schema drift, and enabling self-serve insights with Databricks dashboards and Genie.Time Stamps:00:00 Precaps00:55 Meet the Guest Meghana02:38 Bad Data Real Consequences04:37 Data Engineering vs AI Engineering06:45 Healthcare Example High Stakes09:24 EV Telemetry Scaling Challenges09:58 Synapse vs Databricks Migration14:12 Governance Schema and Photon18:21 Self Serve Analytics with Genie23:21 Technical Debt Before Migration26:32 Future EV Data and V2G29:16 Career Advice Foundations First33:13 Wrap Up and Final ThoughtsConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Meghana Ashok GanatraLinkedin :https://www.linkedin.com/in/meghana-ashok-ganatra/Subscribe to the channel for more conversations with the people building tomorrow's technology.#ai #machinelearning #futureofwork #futureofai #dataengineering #aipodcast #techpodcast #techietalkswithkuldeep data engineering, AI replace jobs, data engineer, data pipeline, bronze silver gold data engineering, data engineering podcast, will AI replace data engineers, EV data analytics, data engineer interview

AI is Overhyped but Underutilized: Last-Mile Enterprise AI, Voice, and India-First Workflows | Techie Talks with Kuldeep | ft. Mathangi Sri Ramachandran, CEO & Co- Founder at YuverseIn this episode , a last-mile AI company building outcome-driven AI solutions for enterprises. They discuss India as a voice-first market, the need for AI to reduce the gap between the “haves and have-nots” in areas like banking and healthcare, and why many AI projects fail due to lack of empathy for end-user workflows rather than technical limits. Mathangi contrasts research papers vs patents, shares how Yuverse applies LLMs, OCR, and agentic orchestration to transform end-to-end enterprise processes like loan and document verification (including messy, handwritten, local-language documents), and explains India-specific conversational nuances for voice bots. She also advises professionals to build breadth—problem-to-outcome skills—over narrow specialization.Time Stamps:0:00 — Trailer / Highlights0:43 — Introduction & Welcome1:27 — Mathangi introduces herself & Yuverse1:42 — Data Science history — from Arthashastra to modern AI4:10 — Top 50 data scientist & enterprise AI adoption4:41 — 100 Patents: Research vs Patents explained5:58 — Building for impact, not just cool tech8:00 — CDO to CEO — Why Mathangi started Uverse9:44 — India's voice-first market & language exclusion problem13:21 — AI for document processing in banking & lending17:40 — Last mile AI & enterprise empathy18:10 — Why 70–80% of AI projects fail21:10 — YubiBERT — India's financial language model22:45 — Future careers in AI — breadth over depth24:57 — Quick Fire Round27:35 — Closing thoughts — AI as transformation, not disruptionConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Mathangi Sri RamachandranMathangi Sri Ramachandran:https://www.linkedin.com/in/mathangisri/Yuverse: https://www.yuverse.ai/Subscribe to the channel for more conversations with the people building tomorrow's technology.#ai #machinelearning #aipodcast #techpodcast #techietalkswithkuldeep

SRE in the Era of AI: Platform Engineering, FinOps, and GenAI for Faster Incident Response | Techie Talks with KuldeepIn this episode, host Kuldeep Singh speaks with Ankesh Khemani, founder of Vibe Monitor and former Atlassian principal engineer, about how Site Reliability Engineering is evolving in the AI era. They discuss why SRE emerged as systems grew too complex with cloud, Kubernetes, and microservices, and how SRE frameworks like SLOs, SLAs, error budgets, and incident management help balance reliability with delivery speed. Anish explains the shift toward platform engineering, expanding responsibilities into cost optimization (FinOps), CI/CD, and internal developer platforms, emphasizing shared ownership with domain teams. They explore how generative AI can analyze code, logs, metrics, and dependency graphs to speed root-cause identification and reduce MTTR, and what new skills matter, including broad systems knowledge, validating AI output, and effective prompting. They also address junior SRE concerns and advise focusing on judgment, architecture, and translating business context into technical direction.Time Stamps:00:00 Highlights01:01 Meet Ankesh Khemani01:28 What SRE Really Is02:49 Why SRE Took Off04:25 Internet Scale Complexity05:28 Platform Engineering Shift07:38 FinOps and Cost Ownership10:01 GenAI Meets Ops11:44 GenAI for Incident Response16:44 New SRE Skills in AI Era20:07 Is Junior SRE Dead22:52 Staying Relevant to 203027:20 Wrap Up and TakeawaysConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Ankesh KhemaniAnkesh Khemani: https://www.linkedin.com/in/ankeshkhemani/Subscribe to the channel for more conversations with the people building tomorrow's technology.#ai #sre #genai #aiindevops #techpodcast #techietalkswithkuldeep

Why In-Person Tech Meetups Matter in 2026 (Stop Lurking, Start Building) The episode argues that despite India’s huge developer base, many remain “invisible” globally because they mostly consume content and lurk in online communities instead of building and showing up in person. Host Kuldeep interviews Monika Rajput, a CNCF Ambassador and senior technical consultant working in data and AI, about how Saturday meetups can create faster technical clarity than long YouTube playlists through real conversations, stories, and direct Q&A. Monika shares how she shifted from over-consuming content to implementing first and then sharing, and how she overcame fear of public speaking after her first unexpected talk at Sharda University. They discuss how community impact—not titles—creates access, stages, and networks, how to choose communities based on aligned tech goals, and practical challenges like finding event space through company partnerships. Monica encourages developers to contribute in any form and speak up before feeling “ready.”Time Stamps00:00 Why India Stays Invisible00:36 Meet Monika Rajput01:55 Stop Consuming Start Building04:31 In Person Beats Virtual07:56 From Back Row to Stage12:28 Ambassador Impact Explained16:31 Choosing the Right Communities20:03 Keeping Meetups Practical21:48 Finding Space and Sponsors25:07 Company Responsibility in OSS27:45 One Action This Month29:20 Anyone Can Build Now31:15 Final Takeaway and OutroConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Monika Rajput Monika Rajput: https://www.linkedin.com/in/mounicarajput/Subscribe to the channel for more conversations with the people building tomorrow's technology.#developercommunity #techmeetups #cncf #opensource #aihackathon #communitybuilding #publicspeaking #AI #artificialintelligence #coding #techpodcast #indiadevs #careerdevelopment #learninpublic #techietalkswithkuldeep

Is It Too Late to Learn AI? Mayur Wadhwani on Switching from Non‑Tech Sales to AI Consulting & Tech SalesIn this episode of Techie Talks with Kuldeep , host Kuldeep Singh interviews Mayur Wadhwani, who transitioned from non-tech sales to AI consulting in sales and solutions at LeewayHertz, helping enterprises identify AI opportunities and translate business problems into practical AI solutions. Mayur explains that moving into tech and AI sales is less about learning to code and more about shifting mindset, understanding business pain points, and learning enough technology to map tools to client needs. He shares how he built tech communities starting in 2016, learned based on client demand, and overcame insecurity through persistent learning. Mayur outlines how top sales and marketing teams use GenAI for real-time workflow drafts, call transcripts and next steps, personalized outreach at scale, lead scoring, deep research, and faster collateral creation. He emphasizes empathy, human judgment, and negotiation as enduring human strengths.Time Stamps:0:00 - Intro0:36 - Welcome & Guest Introduction0:46 - About Mayur Wadhwani1:32 - Is Tech Sales About Coding or Mindset?1:44 - Mayur's Journey: From Logistics to AI Sales9:39 - Do You Need to Be a Programmer to Lead in AI?10:20 - How AI is Transforming Sales & Marketing in 202618:26 - Quick Fire Round18:45 - #1 Skillset Every AI Engineer Needs from Sales19:33 - #2 Most Overrated AI Tool for Marketers20:49 - #3 Will AI Replace Business Development by 2030?23:21 - #4 Biggest Mistake When Switching to Tech/AI26:52 - Key Takeaway & ClosingConnect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Mayur WadhwaniMayur Wadhwani: https://www.linkedin.com/in/mayurwadhwani/Subscribe to the channel for more conversations with the people building tomorrow's technology.#ai #genai #futureofwork #techsales #sales #techietalkswithkuldeep #careertransition #techpodcast #podcast

Is Observability Just Expensive Monitoring? Cutting Costs with Open Telemetry, Grafana & AI Agents (2026)In this episode, Kuldeep talks with architect Mr. Boopesh Sharma about how “observability” has evolved into a costly, overused term and why many teams still treat systems like black boxes despite collecting metrics, logs, and traces. They discuss why “collect everything” has become an expensive failing strategy, and how focusing on relevant telemetry via cardinality analysis, transaction-level monitoring, and correlating metrics, logs, and traces can reduce noise and improve troubleshooting. Bhoopesh explains how AI assistants (e.g., Grafana Assistant, Datadog AI) and agentic workflows are being adopted in production with a human-in-the-loop approach to avoid risky auto-remediation. They also cover OpenTelemetry’s role as a tool-agnostic framework increasingly adopted by vendors, cost-saving practices like using open source in non-prod and SaaS in prod, and the shifting future of the “observability engineer” role toward SRE with AI skills.Time Stamps0:00 — Introduction0:28 — Is observability just monitoring with a 3x price tag?0:45 — How observability has changed: AI, cost & telemetry overload1:01 — People push telemetry data but don't use it — the real pain1:30 — Traditional monitoring vs MELT: metrics, events, logs & traces4:01 — Is "collect everything" a failing strategy in 2026?4:30 — How to filter data and instrument properly with Open Telemetry6:07 — Why your observability bill is higher than your compute bill6:35 — Real-world case: Cutting observability costs by 50% with Grafana9:11 — Is Open Telemetry ready to replace proprietary agents?13:42 — Grafana + Prometheus: Why the open source stack dominates production17:38 — Overrated metrics developers still track (and shouldn't)18:59 — Future of the Observability Engineer job title13:42 — Agentic AI in observability: Human-in-the-loop explained21:14 — Observability for AI: Monitoring LLMs, agents & token costs21:33 — Grafana OpenLIT framework for LLM observability22:33 — Bhoopesh's YouTube journey & open source contributions24:32 — Final takeaways: When your observability cost exceeds business value25:49 — Outro Subscribe for more DevOps, SRE, and Cloud Native content.Connect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Bhoopesh SharmaBhoopesh Sharma: https://www.linkedin.com/in/bhoopesh-sharma/Youtube Channel : https://www.youtube.com/@UCY2gz33KPMbLBuCRtwZFsmw Subscribe to the channel for more conversations with the people building tomorrow's technology.#observability #opentelemetry #grafana #prometheus #devops #sre #aiops #cloudnative #kubernetes #datadog #platformengineering #monitoring #opensource #agenticai #llm #techpodcast #techietalkswithkuldeep

What does it take to file 100+ patents in AI, before the world even knew what AI was ?In this episode, principal scientist and inventor Joy Mustafi discusses his two-decade AI journey from early 2000s neural networks coded from scratch in C to today’s generative and agentic AI. He explains how his passion for mathematics and continuous learning shaped his 100+ AI/ML patents, starting with IBM Watson, and how he identifies problems worth solving by focusing on applied, society-facing domains and bridging academia with industry delivery timelines. Joy compares innovation cultures at IBM, Microsoft, and Salesforce, and shares the origin and growth of Must Research from a small meetup to a nonprofit community focused on AI across domains. He argues modern LLMs still struggle with reliable reasoning and advocates neuro-symbolic, knowledge-graph-based hybrid systems to reduce hallucinations, highlighting a patent on understanding natural-language arithmetic and physics word meanings.00:00 Meet Joy Mustafi00:41 AI Journey Begins02:01 Patent Mindset03:42 Finding Real Problems05:36 Big Tech Cultures07:34 Must Research Origin10:52 Watson to LLMs12:30 Neuro Symbolic Future14:56 Security and Vision16:17 India as AI Creator20:17 Rapid Fire Round21:54 Most Impactful Patent25:37 AGI and Creativity28:31 Fundamentals Over Code30:35 Final Advice Wrap UpConnect with Techie Talks with Kuldeep:YouTube Channel:  @techietalkswithkuldeep  WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about JoyJoy Mustafi :https://www.linkedin.com/in/joymust/MUST Research :https://must.co.in/Subscribe to the channel for more conversations with the people building tomorrow's technology.#ai #artificialintelligence #machinelearning #innovation #patents #ibmwatson #techietalkswithkuldeep #datascience #indiaai

Is Indian Engineering a Bubble? Bridging the Skills Gap with AI, Industry Projects & Real Learning. Every year, India graduates lakhs of engineers. Yet companies still struggle to find people who can actually build, solve, and ship. What's broken — and who's fixing it?The host Kuldeep Singh interviews Dr. Sonal Kukreja, an assistant professor and architect of Spacylabs and Childrenwhocode, about whether India’s engineering boom is a bubble and why companies still can’t find job-ready talent. Sona argues demand for computing is real across sectors, but there’s an oversupply of degrees and an undersupply of capability, especially in tier-2 and tier-3 colleges where education lags industry change. She explains how PEI Labs bridges academia and industry through real projects, constraints, and mentor alignment. The conversation also covers deepfakes and the shift from security to authenticity, the limits of traditional cryptography, and ideas like verifiable media signatures. Sona discusses adapting teaching and evaluation as students use generative AI, emphasizing logic and problem-solving over syntax, and explains why K–12 coding should prioritize conceptual thinking, with approaches tailored for low-resource settings.00:00 Introduction00:31 Is CS a Bubble01:42 Bridging Academia Industry Gap03:51 Tier Two Three Focus06:07 Deepfakes and Authenticity08:21 AI Cheating in Assignments10:57 Kids Coding in AI Era14:46 Village vs Metro Learning18:02 Quick Fire Truth Bombs20:43 Regrets and Career Lessons24:08 Meetups and Real Exposure27:07 Closing Thoughts and ThanksIf you're a student, parent, or educator trying to figure out what actually matters in 2026 — this one's for you.Connect with Techie Talks with Kuldeep:YouTube Channel: @techietalkswithkuldeep WhatsApp Channel: https://www.whatsapp.com/channel/0029Vb7Q4JHAjPXUt8hBgm2TLinkedIn: https://www.linkedin.com/company/techietalkswithkuldeepFacebook: https://www.facebook.com/profile.php?id=61574642927780Instagram: https://www.instagram.com/techietalkswithkuldeepLearn more about Dr. Sonal KukrejaDr. Sonal Kukreja :https://www.linkedin.com/in/drsonalkukreja/Spacylabs :https://www.linkedin.com/company/spacylabs/Childrenwhocode : https://www.linkedin.com/company/childrenwhocode/#engineeringeducation#aiinindia#futureofwork#indianstartups#coding#techietalkswithkuldeep