About
I’m an engineer turned 2x founder who’s spent over a decade building B2B products. Today,…
Articles by Sachin
Activity
30K followers
-
Sachin Gupta shared thisKathryn B. it's been an absolute delight working with you and Beautiful.ai Thank you for being an amazing customer 🙏Sachin Gupta shared thisBeautiful.ai runs an AE-only sales team. No SDRs. So Breakout became their top of funnel. Spotting the companies visiting the site, catching the "just a quick question" inbound, and handing AEs real conversations right when it mattered. The results speak for themselves: - $36,000 generated in pipeline in the trial period - ROI proven in under 6 months Thanks for the trust, Kathryn B. (and the very kind words) 💜 Full story in the comments 👇
-
Sachin Gupta posted thisSalesforce announced #Claudeforce to kickoff the earning call, with Benioff and Dario on CNBC. It's a masterclass on how to sway the Wall Street. But it's a stunning turn when you think about where they were over an year ago. Agentforce came with all the noise about small models trained specifically for task execution. Now they are betting on going headless. They effectively failed to make Agentforce work, but the market rewards a Claude MCP 😆 Honestly, I think this was inevitable. The Claude/Codex harness is just too good. Asking Claude about your pipeline is effortless compared to waiting for a Salesforce dashboard to load and then realizing it doesn't show you the cut you actually wanted. But there's a catch that I don't think anyone is talking about. Salesforce is a genuinely complex product. Exposing MCP for the core data models is the easy part: create an opp, read account attributes, update a stage. But their typical enterprise customer isn't going to get real value from that alone. Enterprises have custom data models, specific workflows, and such. A large surface area of the product has to be exposed through the MCP layer before this feels like anything more than a nicer query interface. Every API needs to be documented in a manner such that a MCP server can make it available. We've been building an AI harness for marketing teams at Breakout, and the thing we ran into is that everything slows down and gets expensive the moment you blow up the tool set in the MCP kit. More tools means more context, more exploration, more tokens, worse latency. It's not a wrapper problem, it's a dedicated engineering problem, and you have to solve it constantly as you add surface area. So the real question isn't whether Salesforce can do it. It's how deep will they go, and whether they'll do it fast enough to make the experience sticky. The part I don't fully understand is the uproar. They exposed write access through MCP and they honor existing permissions through Claude. That's not a leap forward, that's table stakes for anyone shipping an agent on top of a system of record in 2026. Curious if others see something in this I'm missing.
-
Sachin Gupta shared thisA year ago, we staged a funeral for outbound right outside HubSpot's Inbound conference in San Francisco. Hearse circling the Moscone, mourners, recital: the whole nine yards. Sounded crazy on paper, but it turned into our best-performing campaign to date. When every channel is flooded with slop, the alpha is in the unscalable high-effort things no one is willing to do. To close out the year at Breakout, we're lining up a few of these ideas sourced from the team (+ general internet lore) Which brings me to FiDi this morning. Today, the Breakout team is out and about, checking the first one off the list: 📁 unscalable ideas └📁 1. dude-with-sign If your office is here and you're on our list, you might spot some of us holding signs with your names on them. Look outside -- that's us. Come say hi.
-
Sachin Gupta shared thisI tell everyone that cold outbound is dead. Then a cold email I wrote turned into one of our best customers. The reason: timing. Stéphane Le Mentec, LL.M. who runs demand gen at Ketch was using Drift at the time. And like with a lot of chat agents of that generation… the math wasn't working anymore. "This was getting expensive for the results and the ROI that we're getting." (Drift had just had other incidents too. ) He had just started looking around. Ketch sells privacy software to five different buyers. Marketers. Engineers. Security. Privacy. Legal. All five land on the same website, wanting something different. What actually won the evaluation wasn't Breakout's AI or that we out-featured Drift, a Salesloft company or Qualified (Acquired by Salesforce). What pushed us over the line was letting each of those five personas choose their own path. Ketch runs a separate play using Breakout for all five personas landing on their site. The results: → 150+ leads in < 3 mths → 100K+ accounts revealed → 15 min from site visit to first touch (speed to lead) → ~50/50 split between AI-led and live-chat-led leads Thank you to Stéphane Le Mentec, LL.M. and the Ketch team, you've shaped this product as much as you've used it 🙏 Full story in comments.
-
Sachin Gupta posted thisThe deal we just closed at Breakout came inbound for a feature that was a side project a year ago. Last year, I decided that we should spend time building out features that were adjacent to our chat agent: calendar scheduling, routing, and personalized nurtures. From the outside, and probably to our advisors and investors as well, it looked like we were doing too much and losing focus. Fast forward to 2 months ago - an inbound lead shows up whose main need has nothing to do with our main product. Their scheduling tool wasn't cutting it and they were shopping around for a replacement, which is how they found us. They were using Chili Piper, where the scheduling worked but the chat was so weak they'd stopped using it for anything beyond booking. Thanks to the time we spent last year, our product does comprehensive scheduling AND puts an AI agent on their site to handle most of their visitor engagement. We don't have every feature that dedicated scheduling tools have (we check about 70% of the boxes), but we have the ones that our ICP cares most about. I saw the same story play out when we were being evaluated against Qualified (Acquired by Salesforce) in another deal. There’s only so much you can differentiate between one chat agent from another. But we have personalized nurtures for inbound leads, which is painful to run on their platform. “Getting distracted” and building a compound product is proving itself.
-
Sachin Gupta shared thisFor the longest the challenge with enterprise software has been the more feature rich they become the harder they get to use. Configuring settings, pulling data for reports, best practices setup etc. get hard to discover and harder to use. As we are adding more and more capabilities for our customers, we didn't want to fall in the same trap. Over the last 2 months we have invested in building Spoc — an agent that understands everything about Breakout and read and write across our entire data model and features. - Don't see a slice of metric that you are looking for, just ask Spoc - Want best practices on how to set up nudges, just ask Spoc - Want to set up a routing workflow, just ask Spoc Want idea for your next marketing campaign, even for this just ask Spoc 🖖
-
Sachin Gupta shared thisAs a startup founder, you're lucky to get funding. You're luckier when an investor opens doors that turn into customers. But you're the luckiest when they help you solve your toughest problems. That is who Somrat Niyogi is for me. Somrat is an operator first and an investor later. He was at Gusto, founded + exited a few companies. Many investors do not have operator experience, and for those that do, even fewer have his success record. He comes from the GTM world and is deeply embedded. Whenever we want to reach out to someone for Breakout GTM, he usually has a two-degree connection to them. A lot of our early customers came via his introductions. I appreciate how he helps me think. We meet every 2 weeks, and I come in with a specific challenge so we can duke it out and solve it within the hour. He has sharp perspectives and he speaks in specifics. He’s almost never hedging. I can handle strong opinions, and I don't have the time to decipher advice. I'd rather someone tell it to me straight, and Somrat does that consistently. I value when investors go beyond writing a check. It feels like a must-have especially when you're early stage. Somrat, thank you for the work you put in with us.
-
Sachin Gupta shared thisWe hosted a CMO dinner with Storylane last week and the conversation went places. Within about an hour we covered universal basic income, how much should you be putting in 529s, whether our kids will even need the kind of education we got, and what actually happens in the US if AI takes out a large chunk of jobs with no social net underneath. There were some strong opinions, but with humility that nobody can predict one year out, let alone a decade. Then almost in the same breath, the conversation turned to something much smaller and much more immediate: how frustrating it has become to consume AI generated work. Someone described a pattern I think most people in that room recognized. A junior person brings an idea to a meeting. It's well structured, articulate, looks thought through on the surface. You ask one follow up question and it falls apart, because they didn't really think it, they had outsourced their thinking to Claude. The volume of ideas has gone up but the conviction behind them has gone down. That contrast stayed with me on the drive home. At the macro level we're worried AI is going to be capable enough to reorder society. At the table level we're worried it's producing work people can defend. Both are true at the same time. I don't have a clean answer here. I'm certainly not going to tell my team to use less AI, we use it everywhere and it's made us much faster. But I've instituted a rule in the company: you have to be able to defend every line in your doc, if you can't and it's just something Claude wrote, you are wasting everyone's time. The sticky toffee pudding had no such issues. Easily the least controversial thing on the agenda Thanks to everyone who came, great group and a better conversation than I bargained for. Curious whether others are seeing the same thing with their teams. Ruoting Sheel Barbara Nalin Alex Steve
Sachin Gupta replied to a comment
5h
Blake thanks for the mention! 🙏
-
Sachin Gupta liked thisSachin Gupta liked thisI’m growing my team! I’m hiring for two roles across EMEA and APAC at Cast AI: 🇪🇺 Field & Partner Marketing Manager – EMEA: https://lnkd.in/gd8kDiTF 🇮🇳 Senior Field & Partner Marketing Manager – APAC: https://lnkd.in/eHJvJpEV I’m looking for people who are hands-on, creative, commercially minded, and excited about building programs that make a real impact — from Field and Partner Marketing to ABM. If you love working closely with Sales, building relationships, and turning great ideas into pipeline and growth, I’d love to hear from you. Know someone great? Please tag them or share! 🙌 #Hiring #FieldMarketing #PartnerMarketing #ABM #EMEA #APAC Sylvain TACHOIRES | Aditya Nanda | Akash Sharma | Vinayak Bhandarkar | Antoine Aubrun | Ioana Gramschi
-
Sachin Gupta liked thisSachin Gupta liked thisWe just wrapped up 1H here at CaptivateIQ, and before we really get too deep into the next quarter (we are already off to the races! 😆 )- I needed to take a beat and reflect. 1H had its share of changing priorities and new challenges. That's the reality at a company that's growing. Looking back, there are a handful of moments I'm especially proud of 🌟 - We celebrated multiple BDRs who hit (and exceeded) quota (Rayna Cason, Sreya Adusumilli, Max Coleman, William Archer, Samantha Wlodawsky, Kevin Ha, Amanda Pattee) - Continued growing the BDR organization with exceptional new talent. - Cheered on Kristi Bordner and the Enterprise team as they continued to raise the bar, exceeding plan while building consistency quarter after quarter. - Welcomed Kelsie Edmonds to the BD leadership team. She's made an immediate impact, and we're a stronger team because of it. - Hosted our Women in Incentives Breakfast at CaptivateIQ user conference, bringing together an incredible group of women in compensation, including Nicole Redfern & Aya Kawashima. - Connected with customers and fellow operators at both eReward & RevOps Co-op AF, in London. Whether we're hiring, developing talent, launching new partnerships, or spending time with customers, the goal is the same: build an organization that gets better every quarter because the people inside it are getting better every quarter. Excited to see what the second half of the year has in store. 🚀
-
Sachin Gupta liked thisSachin Gupta liked this💙 A Heartfelt Thank You to Team Teal💙 After 7 remarkable years, I have decided to leave Tealium and begin a new chapter. This wasn’t an easy decision but I am proud of what has been built and excited to lean into some new projects. Leading marketing for a company defining a category is a rare privilege. Together, we sharpened Tealium's position as the leader in real-time customer data, elevated the brand on a global stage, and made the case — through story, data, and several very persuasive slides — for why trusted customer data is a strategic advantage, whether powering AI initiatives or relevant customer experiences. Partnering with analysts, customers, and ecosystem partners like AWS, Databricks, and Snowflake, among many others, was an honor. To the executive leadership team — Jeff Lunsford, Ted Purcell, Ali Behnam, Mike Anderson, Doug Lindroth, Phil Maynard, Dave Mueller, and Adrianne Court — thank you for your vision, partnership, and belief in what marketing can drive when it's given a real seat at the table. Your trust made everything possible. 🙏 To the global marketing team — current and alumni alike — thank you for your creativity, rigor, and relentless commitment to excellence. Together we built a marketing engine that didn't just tell Tealium's story; it moved the business. Brand, demand, and pipeline all up and to the right — because of you. 🌎 And 12 out of 10 marketers surveyed agree: after 25+ years of building brands, teams, and pipeline, a little white space is allowed. 😄 To everyone at Team Tealium — thank you for the collaboration, innovation, and countless memories along the way. I'll always be proud of what we accomplished, and I'll be cheering you on from the sidelines — loudly, and probably with a dashboard open. 💪 Onward and upward — the best is yet to come! ✨
-
Sachin Gupta liked thisSachin Gupta liked thisWe are hiring: Marketing Enthusiast at Relific I am looking for someone genuinely excited about marketing and eager to learn by doing. At Relific, things are not always neatly structured. We are building, experimenting, figuring things out, and sometimes changing direction. So, I am not looking for someone who needs a perfectly defined role or a checklist of things to do every day. I am looking for someone who is: • Curious and willing to learn • Comfortable taking up new challenges • Good at communication and content • Willing to experiment and try new things • Comfortable working with ambiguity • Someone who takes ownership and gets things done You don't need to know everything about marketing. What matters more to me is the willingness to learn, take responsibility, and figure things out. I am willing to give the right person the time, space, and guidance to learn and grow. You will get exposure to content, branding, social media, customer stories, market research, campaigns, sales enablement, and many things that may not even have a name yet. If you are someone who thinks: “I don't know this yet, but I can figure it out.” — I would like to hear from you. Location: Bengaluru / Hybrid Experience: 0–3 years If this sounds like you, DM me or share this with someone who might be a good fit. #Hiring #Marketing #StartupJobs #Relific
-
Sachin Gupta liked thisWe put our AI SDR inside Zoom because manually booking sales meetings in 2026 feels like using dial-up. ⚡ https://lnkd.in/d8Qs8kz4Sachin Gupta liked thisWE'RE ON THE ZOOM MARKETPLACE our AI SDR now lives inside Zoom and books meetings for you. on platform. off platform. it does not care where the meeting is. it just wants the meeting to happen. it's there. booking things. while you do literally anything else. check it out - https://lnkd.in/d8Qs8kz4
Experience & Education
Publications
-
Efficient Variable Size Template matching Using Fast Normalized Cross Correlation on Multicore Processors
LNCS Springer
Normalized Cross Correlation (NCC) is an efficient and robust way for finding the location of a template in given image. However NCC is computationally expensive. Fast normalized cross correlation (FNCC) makes use of pre-computed sum-tables to improve the computational efficiency of NCC. In this paper we propose a strategy for parallel implementation of FNCC algorithm using NVIDIA’s Compute Unified Device Architecture (CUDA) for real-time template matching. We also present an approach to make…
Normalized Cross Correlation (NCC) is an efficient and robust way for finding the location of a template in given image. However NCC is computationally expensive. Fast normalized cross correlation (FNCC) makes use of pre-computed sum-tables to improve the computational efficiency of NCC. In this paper we propose a strategy for parallel implementation of FNCC algorithm using NVIDIA’s Compute Unified Device Architecture (CUDA) for real-time template matching. We also present an approach to make proposed method adaptable to variable size templates which is an important challenge to tackle. Efficient parallelization strategies adopted for pre-computing sum-tables and extracting data parallelism by dividing the image into series of blocks substantially reduce required computational time. We show that by optimal utilization different memories available on CUDA and using idling time of host CPU to perform independent tasks we can obtain the speedup of the order of 17X as compared to the sequential implementation.
Other authorsSee publication -
Motion Detection in Low Resolution Grayscale Videos Using Fast Normalized Cross Correrelation on GP-GPU
ICAISC, Bhuvaneshwar
Motion estimation (ME) has been widely used in many computer vision applications, such as object tracking, object detection, pattern recognition and video compression. The most popular block based similarity measures are the sum of absolute differences (SAD), the sum of squared differences (SSD) and the normalized cross correlation (NCC). Similarity measure obtained using NCC is more robust under varying illumination changes as compared to SAD and SSD. However NCC is computationally expensive…
Motion estimation (ME) has been widely used in many computer vision applications, such as object tracking, object detection, pattern recognition and video compression. The most popular block based similarity measures are the sum of absolute differences (SAD), the sum of squared differences (SSD) and the normalized cross correlation (NCC). Similarity measure obtained using NCC is more robust under varying illumination changes as compared to SAD and SSD. However NCC is computationally expensive and application of NCC using full or exhaustive search method further increases required computational time. Relatively efficient way of calculating the NCC is to pre-compute sum-tables to perform the normalization referred to as fast NCC (FCC). In this paper we propose real time implementation of full search FCC algorithm applied to gray scale videos using NVIDIA’s Compute Unified Device Architecture (CUDA). We present fine-grained optimization techniques for fully exploiting computational capacity of CUDA. Novel parallelization strategies adopted for extracting data parallelism substantially reduce computational time of exhaustive FCC. We show that by efficient utilization of global, shared and texture memories available on CUDA, we can obtain the speedup of the order of 10x as compared to the sequential implementation of FCC.
Other authorsSee publication
Courses
-
Compilers
-
-
Database Management Systems
-
-
Operating System
-
-
Operating System
-
Honors & Awards
-
Forbes 30 under 30
Forbes
Awarded as Forbes 30 under 30 in the Enterprise Tech category for Asia.
-
Forbes 30 under 30
Forbes
Recognized in Forbes 30 under 30 for Enterprise software.
Languages
-
English
Native or bilingual proficiency
-
Hindi
Native or bilingual proficiency
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content