Swapan Rajdev

While it's annoying, it's a bullet dodged. Such basic courtesy and communication skills speaks a lot about who they are. I've also experienced the other way round, interviewers are a no-show without any prior communication. Very frustrating but speaks a lot about the company culture.

I'm sort of baffled. I've been spending 30 hours a week interviewing strategists & engineers. Probably 20% of candidates never show up for the call. Is this a generational thing? Is it just that people are poor communicators? Imo, nothing more entitled than no showing to a job interview.

Measuring the quality of Voice AI agents is messy. You have to score it on what was said, how well it was said, how fast it responded, and whether it solved the task at hand. And all of this with no ground truth!

Signal Dialogues #01 is live. We built Signal because there wasn’t a clear, consistent voice telling the story of AI in India — across founders, capital, research, and policy. For our first episode, @aakrit sits down with @vkhosla and @mukundjha to go over the rise of Emergent — probably the fastest growing software company in history to hit $100M ARR. Shot at the sidelines of the @OfficialINDIAAI summit at the iconic @iitdelhi. Timestamps: 00:00 – Intro: Fastest Growing Software Company Ever? 07:28 – Mukund's Journey: Google → Dunzo → Emergent 22:23 – You're Limited by What You Think You Can Do 27:08 – The $3M Bet That Built the Internet 47:11 – $100B AI Company From India?

Ever since we started building Activate, I knew we would eventually launch a podcast. But it couldn’t be just another VC show. While there is tremendous energy around AI in India today, there isn’t yet a clear, consistent platform bringing together founders, capital, research & policy in one conversation. So today, we’re launching Signal Dialogues. For our very first episode, I couldn’t have dreamt of a better setting: a legend, India's AI poster boy, an iconic venue, and at the country's most defining AI moment. Privilege to have none other than @vkhosla together with @mukundjha, co-founder of @emergentlabs, one of the world's fastest ever software companies to hit $100M ARR, on the sidelines of the @OfficialINDIAai Impact Summit at @iitdelhi. #DontStop 🚀

Best time, best team! Congratulations and God speed! @aakrit @177pc

>> git checkout -b dashboard_cleanup >> claude >> {explain all the clean-ups required} >> Yes, allow all edits in this session Power nap! My current state, happy place :)

@KaranVaidya6 @composio @lightspeedvp Congratulations @KaranVaidya6 and @GanatraSoham. Top Video 👍

Link to the full post below: srajdev.com/p/llm-evaluati…

I've seen this pattern in every AI app I've worked on. Demos work great, but then real users expose the cracks. Change one prompt and something else breaks. Users and clients start losing faith in AI. The answer to most of these problems? - Strong, clear evaluation metrics that you track regularly. - It gives you the confidence and peace of mind of knowing exactly how your app is performing in production. I wrote a guide on the scorecard I use for all the apps I ship and how to implement it. Link to the full implementation guide in the comments.

@PrajwalTomar_ nike

a lot of you asked how i built this hero section using lovable, midjourney, and runway → comment “nike” and follow i’ll dm you the exact steps, prompts, assets, and the remix link

I went down a bit of a rabbit hole recently trying to understand AI guardrails. What started as “how do I stop bad outputs?” turned into a deep dive into architecture, safety frameworks, unintended behaviors, and a dozen real-world incidents where things almost went off the rails. Most people treat guardrails like seatbelts: a safety add-on. But in agent systems that plan, reason, call tools, and store memory, guardrails have to be part of the architecture, not just wrappers around LLM outputs. I pulled everything I was learning into a blog post and realized that before talking about how guardrails work, we need to understand what exactly we're trying to protect. 🧩 Here are the 6 key attributes that good guardrails should defend: - Accuracy: prevent hallucinations & disinformation - Privacy: avoid sensitive data leakage - Security: block injections, unsafe code/tool access - Safety: prevent harmful or dangerous behavior - Compliance: align with legal/regulatory standards - Fairness: reduce bias in reasoning and responses Read the full post to explore how guardrails work and why thinking of them as runtime actions across your entire agent stack is a much more powerful model than just filtering input and output. Link to the full post here: srajdev.com/p/multi-layere…

Simple stuff like this makes me productive and a bit excited because I feel like I just found a smarter way to work :) Told ChatGPT to go through all the references in a research paper I am currently studying and turn them into a table with links. No more googling each one manually. Note: had to tell it to scrape 10-20 at a time, then ask it to merge them all.

If you want your team actually to use AI and not just talk about it - here’s an interesting idea: 👇 Lenny Rachitsky just dropped the AI Bundle of the Year — $14,000+ worth of top AI tools (Cursor, Replit, Notion AI, Perplexity Pro, and more) for just $200: lennysnewsletter.com/p/an-unbelieva… Here's how to turn this into magic inside your org: ➡️ Run a 10–15 day challenge where employees share their most creative use of AI for personal growth or productivity ➡️ Reward the top 3–5 entries with this bundle (gifted to their email so they keep it) ➡️ Watch how fast AI adoption takes off It’s recognition + upskilling + culture-building in one shot. Best $200-$1000 you will spend this year.

Link to the post: srajdev.com/p/a-deep-dive-…

Pretraining gives language models their knowledge, and post-training gives them purpose! But with so many techniques—fine-tuning, instruction tuning, RLHF, test-time reasoning—it’s easy to get lost in the maze. I recently came across a fantastic paper—LLM Post-Training: A Deep Dive into Reasoning LLMs—that mapped out this space beautifully. So I decided to break it all down into something easier to digest. 📘 In this post, I’ve: - Categorized key post-training methods (Fine-Tuning, RL, Test-Time Scaling) - Summarized strengths, challenges, and real-world use cases - Shared clean, visual tables to help you pick the right technique for your application This reference might save you a lot of time if you’re building AI products or working with LLMs in production.

Reading a good book is like binge-watching a show, you find every excuse to procrastinate work. The only difference? You don’t feel as guilty because reading feels productive.

🚀 In the AI era, being a 'wrapper company' isn't a limitation—it's an opportunity! With foundational models now widely accessible, the true differentiators lie in:​ - Speed: Rapid iteration and agile development to stay ahead.​ - Quality: Delivering exceptional user experiences that resonate.​ - Distribution: Crafting innovative go-to-market strategies and building strong communities.​ Embracing the 'wrapper' approach means focusing on:​ - AI-native workflows: Reimagining processes to fully leverage AI's potential.​ - User-centric design: Creating interfaces that simplify complexity.​ In this landscape, it's not just about the technology but how you apply it. The companies that thrive will be those that integrate AI thoughtfully, enhancing value and redefining industries.​ You can read more in my latest post below.

For 30 years, I’ve witnessed the power of a great special educator—because my mother is one. The right guidance doesn’t just help children learn; it transforms entire families, especially for differently abled children. But in India, the reality is stark: 📌 30 million children are estimated to have Specific Learning Disabilities (SLD). 📌 Yet, only 8 million have been officially identified—leaving 22 million children and families unaware. 📌 Even for the identified cases, we have just 1.6 lakh special educators—meeting only 25-35% of the actual need. This massive gap in special education is a challenge—but also an opportunity. Given where AI stands today, I truly believe we can bridge this gap and create a lasting impact. That’s why I started building an AI-powered special educator—to bring personalized, accessible, and scalable support to children, parents, and educators. 🚀 Super excited to share that on the 1-year anniversary of IndiaAI Mission, we were selected as one of 30 companies (from 900+ applicants) to receive a grant from the IndiaAI Application Innovation Challenge. This recognition strengthens our mission to revolutionize special education in India. If this vision resonates with you, I’d love to connect. Whether you're an educator, technologist, policymaker, or just someone who cares deeply about making a difference—we need you. Let’s build the future of special education together. #AIForGood #InclusiveEducation #SpecialEducation #IndiaAI #Accessibility