About
I have worked across roles, consulting, product management and marketing, business…
Articles by Rajesh
-
Serendipity, Second gig and Software - how it all converged at Devtron
Serendipity, Second gig and Software - how it all converged at Devtron
Software continues to eat the world ! Post mCarbon's acquisition to Nuance, While I was navigating a transition and…
20 Comments -
Its time you start the love affair
Its time you start the love affair
The “Aha” moment to start The good thing is that you are already thinking about it or have thought about it but not…
10 Comments
Activity
10K followers
-
Rajesh Razdan reposted thisRajesh Razdan reposted thisStartups are hard. Travel startups are brutal. We are looking for the ones who tried anyway. We have deep respect for every founder and early employee who bet on themselves in travel - navigating antiquated API integrations, unreliable suppliers, razor-thin margins, and customers who expect perfection while the world throws Covid, wars, oil shocks, and currency chaos at you. At TravClan, we're building an A-team of such builders and operators not despite what happened, but because of it. You will own a product / business line end-to-end. No politics, full accountability, full ownership. Top-tier comp + meaningful ESOPs. We are at an inflection point with opportunity to expand globally with AI being a sharp lever to scale further. Today we have crossed ₹1,000 crore GMV and are growing fast serving 5,000+ travel agents & OTAs. 50% of our managers are already leading AI projects. We expect you to lead, not observe. To apply, email grow@travclan.com with email subject: Builder at TravClan | <Your name> and include what you built in travel, what you'd do differently and your resume / linkedin profile link. The crazy ones built airlines, OTAs, and tour companies that changed how the world moves. If you are are between chapters right now. This is your chance to add one to your story!
-
Rajesh Razdan reposted thisRajesh Razdan reposted thisJust launched agent-skills: an open-source collection of AI agent skills that teach AI assistants how to write, run & maintain real-world test automation across 15+ frameworks. AI tests smarter, not harder. 🔥 https://lnkd.in/d7yYiJuiGitHub - LambdaTest/agent-skills: Agent Skills for Test AutomationGitHub - LambdaTest/agent-skills: Agent Skills for Test Automation
-
Rajesh Razdan posted thisHeroku mattered. For a long time, it represented something important — speed, focus, and a developer experience that got out of the way. Many great companies were built on it. A quiet shift is happening in developer platforms. The old promise was: “Don’t think about infrastructure.” The new reality is: you can’t avoid it — but you shouldn’t have to suffer through it either. As teams rethink Heroku and similar platforms, I’m seeing a common pattern: -They want hyperscaler power and Kubernetes portability, but they don’t want YAML-driven chaos or a maze of bespoke pipelines. This middle layer — between raw cloud and fully managed PaaS — is where the next generation of platforms will win. 🚀 🚀 That’s the problem space we’re deeply invested in at Devtron Inc., and honestly, it’s the most interesting infrastructure challenge I’ve seen in a decade. Would love to hear how others are approaching this transition — or what’s been harder than expected ?
-
Rajesh Razdan shared thisIt’s about the architectural necessity of Decoupling Deployment from Release. Great read by our colleague Pawan Kumar. There is deployment and then there is “Devtron Deployment” Devtron Inc.Rajesh Razdan shared thisIf "Deployment" and "Release" mean the same thing to you, you don't have a CI/CD pipeline, you have a gamble. In native Kubernetes, when you update a Deployment, you expose that new code to your users immediately. If the new version has a critical bug, you’ve just maxed out your error budget in seconds. This is why high-scale infrastructure eventually needs a Service Mesh. It’s not about the hype. It’s about the architectural necessity of Decoupling Deployment from Release. In the latest edition of The Five Nines Report, I break down why Layer 7 traffic control is the only safe way to scale releases.Why You Actually Need a Service Mesh: Decoupling Deployment from ReleaseWhy You Actually Need a Service Mesh: Decoupling Deployment from Release
-
Rajesh Razdan posted thisMost teams don’t actually have a tooling problem. They have a complexity problem. Over the last few years , we’ve built incredible systems — cloud-native stacks, Kubernetes, CI/CD, AI-assisted development. Each layer solved something real. But together, they’ve created an environment where shipping reliably is as important than shipping fast. At Devtron Inc., what we see repeatedly is this: Teams want confidence, not dashboards. Confidence that a change won’t break production. Confidence that automation is doing the right thing. Confidence that AI suggestions are grounded in their systems, not generic patterns. That’s what our platform is really about today — learning from how software behaves in the real world and using that intelligence to guide delivery, verification, and decision-making. Less guesswork. Fewer rollbacks. More signal, less noise. Where this gets interesting is where the industry is heading next. AI in DevOps and cloud native won’t be about chatbots bolted onto tools. It will be about systems that understand context — your code, your infrastructure, your traffic, your risk tolerance — and can act autonomously within guardrails you trust. That’s the future we’re building toward: AI that doesn’t just recommend, but reasons. Automation that doesn’t just execute, but adapts. Platforms that fade into the background because they’re finally working the way teams think. Curious how others are approaching trust, autonomy, and AI in software delivery. What’s working — and what still feels broken?
-
Rajesh Razdan shared thisToday, we’re excited to introduce Devtron SaaS — the fastest way to deploy applications and manage Kubernetes infrastructure at enterprise scale. Built for platform and DevOps teams who want the power of a unified DevOps platform, without the operational burden of managing the underlying infrastructure.Rajesh Razdan shared thisManaging Kubernetes is hard enough - running the DevOps platform behind it shouldn’t be. Today, we’re excited to introduce Devtron SaaS - the fastest way to deploy applications and manage Kubernetes infrastructure at enterprise scale. Built for platform and DevOps teams who want the power of a unified DevOps platform, without the burden of managing the underlying infrastructure. With Devtron SaaS, you get: ✅ Instant access - start deploying in minutes ✅ No infrastructure setup - skip cluster provisioning entirely ✅ Full enterprise-grade Devtron capabilities - AI Agent, Application Management, Infrastructure Management, FinOps, and more. Focus on building and scaling applications. We’ll take care of the platform. 👇 Sign up today and start your 30-day free trial of Devtron SaaS. #Devtron #Kubernetes #DevOps #PlatformEngineering
-
Rajesh Razdan reposted thisRajesh Razdan reposted thisIn 2025, we shipped several major updates to help teams scale and manage Kubernetes with ease. Here’s a quick recap of key features we released this year! #Devtron
-
Rajesh Razdan reposted thisRajesh Razdan reposted thisFrom all of us at Devtron, Happy Holidays! Reflecting on 2025, we’re inspired by the progress we’ve achieved together with our customers, partners and teams worldwide. Wishing you a joyful holiday season and a bright New Year ahead. #HappyHolidays #Devtron
-
Rajesh Razdan posted thisWe spoke to more than 20 leaders over the past many weeks from Kubecon in Atlanta to AWS re:Invent and everyone talked about them experiencing more complexity to handle both infrastucture and application, given acceleration of code generation nowadays with AI copilots while deployments beginning to feel the bottleneck as one navigates faster time to market and reliability intact, in the same sentence. "This is where the Devtron platform comes in, enabling both development and SRE teams to handle daily production use cases such as deploying new software, debugging production incidents, and monitoring production software. Devtron uses AI to analyze and automate SRE functions" Devtron Inc. 's roadmap to address this is quite unique whether you are migrating from legacy deployments to new or going from modern to a new AI normal. Give it a try at https://devtron.ai/
Rajesh Razdan commented on a post
2w
These are great examples where Systems of records sit and deeply embedding AI in both Supreme court and CAG with IITM's Innovation framework is very exciting to say the least. Way to go ! Gaurav Raina
Rajesh Razdan commented on a post
1mo
The agents need to sit within system of records to get true enterprise value and wealth management is one such nuanced function where context is more critical than template and hence you need a purpose built system not just a model. Era Jain your hypothesis is correct.
-
Rajesh Razdan liked thisThis is a major step toward automatically securing software. If your team is struggling with security backlog, and agentic coding is only adding to your struggles, read this.Rajesh Razdan liked thisHow GitLab's new Agentic SAST Vulnerability Resolution is changing the game. Unlike traditional tools that just flag problems, GitLab Duo functions as a reasoning agent: * Traces vulnerabilities across the entire codebase to understand the root cause. * Generates high-quality, context-aware code fixes (not just single-shot suggestions). * Automates pipeline validation on proposed fixes, before a human review. This isn't just an assist—it's a fundamental paradigm shift toward autonomous security remediation. https://lnkd.in/dNMXZpRe #GitLab #AI #DevSecOps #Cybersecurity #SoftwareDevelopment #SunfireTechnologiesHow GitLab’s Agentic AI is Rewriting the Rules of Vulnerability RemediationHow GitLab’s Agentic AI is Rewriting the Rules of Vulnerability Remediation
-
Rajesh Razdan liked thisRajesh Razdan liked thisA slow Sunday afternoon, looking back at the last year at Nerah Skin™ When I started, all I had was one table, one chair, racks of products, and a question I kept returning to, can skincare be made with clarity, honesty, and fewer steps? The year that followed came with rejections, pauses, and moments of reassurance that we were moving in the right direction. Being bootstrapped meant learning as we went, staying close to the work, and being careful about what we chose to build and what we chose not to. Today, we’re a full-time team of five. As we step into a new financial year in April, this moment feels significant. We’re closing our first year of operations, profitably :) and a clearer sense of direction for what comes next. Year two feels less about experimentation and more about innovation: sharper products, stronger systems, and deeper alignment with our community. Writing took a backseat while I was heads-down finding the right people to build with. Sharing this now feels overdue. #foundernote #notshy #itsmyjourney
-
Rajesh Razdan reacted on thisRajesh Razdan reacted on thisCongrats to Citi Ventures portfolio company Skan AI and CEO Avinash Misra, Co-founder, CPO/COO Manish Garg, and Aman Rangrass, SVP, Global Head of Revenue on the Nasdaq feature, a very well deserved recognition. Skan AI is tackling one of the hardest (and most underappreciated) problems in enterprise AI: agents can’t act effectively if they don’t understand how work actually happens. Their approach, observing real work across systems and building execution-level intelligence, is what enables enterprise-grade agentic automation. Skan’s platform continuously captures “ground-truth execution visibility” across front, middle, and back office workflows, which becomes the foundation for deployable, governed AI agents. This is critical because: • Most agents today are interface layers over LLMs, not systems of execution • Enterprises need process-level context, not just language understanding • Governance, auditability, and ROI all depend on how work actually gets done Skan’s thesis, that enterprise AI needs a system of record for execution, is exactly what bridges that gap. As Avinash Misra puts it, their work effectively connects “human intent and autonomous execution” in a scalable way. Looking forward to hosting Avinash on a panel on Enterprise Agentic AI at the upcoming Citi AI Summit. Should be a great discussion on what it really takes to move from automation to autonomy.
-
Rajesh Razdan liked thisRajesh Razdan liked thisI wrote something personal this week. It starts with a black and white television and ends with building an AI lending agent. In between: a begged entry into a computer lab, four years of sleeping under my desk, a Panasonic phone from a South Delhi basement, and a chat room handle called mouse_potato. Every shift in my career started the same way, playing with something before I understood it. This one is no different. Link in comments If you've ever felt the pull of the next thing before you could explain why, this might resonate.
-
Rajesh Razdan liked thisRajesh Razdan liked thisIf I had to rebuild my entire AI engineering stack from scratch, these are the GitHub repos I'd start with. RAG & Retrieval → https://lnkd.in/gc6iqNzJ — de facto standard for RAG pipelines → https://lnkd.in/gJ8gKkWt — best data ingestion and indexing layer → https://lnkd.in/gZ9rssT4 — most mature enterprise RAG framework → https://lnkd.in/g59pdntX — graph-based RAG, outperforms naive chunking on complex queries Evaluation → https://lnkd.in/g_W6rMPP — CLI for testing prompts, agents, RAGs. Used by OpenAI and Anthropic → https://lnkd.in/ghm2KjCV — pytest-like unit testing for LLM apps → https://lnkd.in/gAPdCRzh — RAG-specific evaluation: faithfulness, relevancy, context precision Agents → https://lnkd.in/gE6uUyFP — fastest-growing multi-agent framework → https://lnkd.in/g8bRBpYZ — graph-based agent orchestration with state management → https://lnkd.in/gFWrZKmt — minimal, model-agnostic agents from HuggingFace → https://lnkd.in/gntUMKN8 — type-safe agents built on Pydantic Vector Databases → https://lnkd.in/g7CzWvVS — Rust-based, best metadata filtering → https://lnkd.in/gnKp9w23 — simplest DX for prototyping → https://lnkd.in/gfQeQrrx — foundational similarity search, many vector DBs use it internally Fine-Tuning → https://lnkd.in/gGthsfTa — 2-5x faster fine-tuning, 80% less memory → https://lnkd.in/g5hwqBZN — unified framework for 100+ models with web UI → https://lnkd.in/gzyc_SQ8 — official HuggingFace library for SFT, RLHF, DPO Production & Deployment → https://lnkd.in/gS3WT5wj — run LLMs locally with one command → https://lnkd.in/gK5DSbKp — default for production LLM serving → https://lnkd.in/gtpYuDP5 — rising star, 4.6x faster than vLLM in benchmarks → https://lnkd.in/gyyH_gEX — the go-to framework for serving AI applications as APIs Observability → https://lnkd.in/gS9QB3bM — open-source LLM tracing, evals, prompt management → https://lnkd.in/g9wPBReC — AI observability with OpenTelemetry-native tracing → https://lnkd.in/gWkztfUX — open-source LLM evaluation and tracing platform Structured Output → https://lnkd.in/gPDWwSf8 — the standard for data validation in Python, foundation for structured LLM outputs → https://lnkd.in/gp5tT6b2 — constrained generation from LLMs → https://lnkd.in/gh4tBbeT — guaranteed valid JSON, regex, grammar output Save this. Pick the category you need and start there. _____ 👋 P.S. If you want to learn how to actually use these in production, not just star them, the Engineer's RAG Accelerator covers RAG, evaluation, agents, and deployment hands-on. April cohort → [Visit my website] ♻️ Repost to help an engineer find the right tools.
-
Rajesh Razdan reacted on thisRajesh Razdan reacted on thisIncredibly proud dad moment this week as my 14 year old daughter (artist name ‘Smy’) releases her debut single ‘sleepless’. Just in time for our weekend playlists! #smy #smyrakamity #sleepless https://lnkd.in/gNHcWrbK
-
Rajesh Razdan liked thisExciting launch! Been using Dimension for a month and it saves me so much time! Give it a try! Love the daily morning briefs and the email auto responses! Tejas RavishankarRajesh Razdan liked thisIntroducing Dimension. The AI coworker that never sleeps. Dimension plugs into your entire stack and gets to work. Gmail, Calendar, Slack, GitHub, Linear, Notion, and 30+ apps. Morning briefings. Meeting prep. Email drafts. Daily recaps. All handled. You can chat with it on Slack, iMessage, or the web. Set it up in 2 minutes. Assign Dimension to the busywork so you can spend your time on what actually matters. Get started at dimension.dev
-
Rajesh Razdan liked thisRajesh Razdan liked thisJust wrapped an incredible conversation with some of the brightest minds in AI. Here are 5 things that might shape the next 2-3 years: 💡 Physical sciences are next. AI is moving beyond text and code into materials science, chemistry, and mechanical engineering. Research labs and industry players are already raising billions to capitalize on this shift. 💡 Enterprise data is the real moat. Large CRM companies already have hundreds of vetted, security-tested agents ready to deploy. Startups without access to enterprise data pipelines simply cannot compete and the window is closing fast. 💡 Orchestration layer startups are in trouble. Major SaaS platforms are building their own orchestration layers in-house. Unless you're in a niche vertical or positioned for acquisition, the business case for third-party orchestration is disappearing. 💡 Productivity gains are already dramatic. Product leaders are seeing real 4x output increases through agent automation - SDLC cycles compressed from months to weeks, product managers submitting PRs, and entire workflows running autonomously in the background. 💡 Knowledge democratization is reaching its ceiling. Everyone now has access to the same information base. The next frontier isn't consuming existing knowledge but generating new knowledge, where formal methods + LLMs will drive the next wave of breakthroughs. I am curious to hear your thoughts! Are there other trends you are paying close attention to? #ai #worldmodels #vc Seligman Umesh Eddie
-
Rajesh Razdan reacted on thisRajesh Razdan reacted on thisStartups are hard. Travel startups are brutal. We are looking for the ones who tried anyway. We have deep respect for every founder and early employee who bet on themselves in travel - navigating antiquated API integrations, unreliable suppliers, razor-thin margins, and customers who expect perfection while the world throws Covid, wars, oil shocks, and currency chaos at you. At TravClan, we're building an A-team of such builders and operators not despite what happened, but because of it. You will own a product / business line end-to-end. No politics, full accountability, full ownership. Top-tier comp + meaningful ESOPs. We are at an inflection point with opportunity to expand globally with AI being a sharp lever to scale further. Today we have crossed ₹1,000 crore GMV and are growing fast serving 5,000+ travel agents & OTAs. 50% of our managers are already leading AI projects. We expect you to lead, not observe. To apply, email grow@travclan.com with email subject: Builder at TravClan | <Your name> and include what you built in travel, what you'd do differently and your resume / linkedin profile link. The crazy ones built airlines, OTAs, and tour companies that changed how the world moves. If you are are between chapters right now. This is your chance to add one to your story!
Experience & Education
-
Devtron Inc.
View Rajesh’s full experience
See their title, tenure and more.
or
Languages
-
German ( basic )
Elementary proficiency
-
English
Full professional proficiency
-
Hindi
Professional working proficiency
-
Spanish
Elementary proficiency
Recommendations received
1 person has recommended Rajesh
Join now to viewOther similar profiles
-
Patrick Veron
Patrick Veron
Highly accomplished customer-focused leader with an extensive experience in leading sales teams, articulating and executing strategic and business development plans or sales strategies that are delivering actual growth for the business. <br><br>Delivered hyper-growth of up to 60% YoY in different lifecycles, managed teams of up to 350 people, and led sales teams across regions (NorAm, EUR, APAC). The yearly sales have consistently been in the 9 numbers (USD). I therefore have a strong track-record in building, leading and coaching teams across Asia Pacific, Europe and Americas in business development, sales, marketing and overall customer success. <br><br>Strategic, hands-on, entrepreneurial, analytical with a problem-solver approach, I am focused on delivering growth and sales leadership, supporting my team members to develop relevant value propositions for existing and new customers.
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Rajesh Razdan
14 others named Rajesh Razdan are on LinkedIn
See others named Rajesh Razdan