About
Living deliberately and immediately. Passionate about building products that make a…
Articles by Sanket
Activity
7K followers
-
Sanket S. shared thisSometimes you come across an example of sheer ambition and agency, and it hits you like a brick. Today was one of those days, watching the Midjourney Medical announcement and an early demo of their full body scanner. Long SF (whose much-discussed demise was greatly exaggerated).
-
Sanket S. shared thisAmazing work by the team in getting SCA support landed for monorepos!Sanket S. shared thisIf you run DeepSource in monorepo mode, today's release is for you. Monorepo mode lets you map subfolders of a single git repo as first-class repos on DeepSource, each with its own settings, dashboards, and configs. Until today, the one piece missing was SCA. That's fixed. Every sub-repo now gets its own Dependencies tab, with vulnerabilities scoped to the packages that sub-project actually uses. No noise from elsewhere in the monorepo. Clean ownership, clean triage. Two more updates in this release: AI Review just opened up to 10 more languages: Dart, Elixir, Apex, Groovy, Objective-C, VB.NET, PowerShell, Lua, Erlang, and Perl. Your PRs in these languages now get the full DeepSource review treatment. And for self-hosted teams: Enterprise Server now supports the Kubernetes Gateway API. If you've moved off Ingress, DeepSource fits in natively. Full notes: https://lnkd.in/gkE5pN_cOSS vulnerability scanning for monorepos • DeepSource ChangelogOSS vulnerability scanning for monorepos • DeepSource Changelog
-
Sanket S. shared thisAs AI models get better, those who build the best harness and context engine around the model will reap the most benefits. True for coding agents (why Codex / Claude Code are astronomically better than the competition), true for AI code reviewers. Our team's been cooking hard lately. In addition to switching the underlying models, we did several foundational improvements to our static analysis-based hybrid harness for this release. As a result, we're seeing considerable gains on all key metrics. This means we're able to find more vulnerabilities in our customers' code with fewer false positives.Sanket S. shared thisDeepSource AI Review now scores higher on the OpenSSF CVE Benchmark. Most importantly, we're seeing 4%+ gain on the F1 score. This means we're catching more real vulnerabilities without introducing a single false positive. Our precision hit 100% (zero false positives across 83 patched diffs) while recall climbed to 73.17%. Read our full results and methodology: https://lnkd.in/gXG-KZha
-
Sanket S. posted thisWriting is often where the thinking happens. When I’m drafting a memo, PRD, or strategy doc, the real value is usually in finding the parts that don’t hold up yet. Writing exposes fuzzy thinking fast. That’s why handing the argument to AI too early is risky. It can make weak ideas sound complete before they’re actually solid. I use AI constantly for research, cleanup, and pushback. But I still want to do the core writing myself. If the draft comes out too clean, too early, that’s usually a warning sign.
-
Sanket S. shared thisOne of my favorite DeepSource + Claude Code loops is autonomously fixing CVEs in our projects with our MCP Server (that just went out live today). Simple workflow: > ask Claude Code (or any other agent): ''review vulnerabilities in this repo and upgrade packages" > the agent uses DeepSource MCP to get all vulnerabilities detected in the repo with complete CVE information and DeepSource's inferred reachability data > after filtering out unreachable and low priority vulns., the agent uses the "fixed in" versions for each package and upgrades them > one single patch with all vulnerabilities fixed (see video below) This is rolling out to all our customers today. We're putting code security on autopilot. The future is agentic!
-
Sanket S. shared thisWhen Aditya showed me the first prototype of what they're building (it wasn't called Noon then!), I knew they were on to something big. Agentic coding has irrevocably changed how we build software. The time for product design has come. Thanks for letting me be a small part of the story and all the best Aditya and Kushagra! <3Sanket S. shared thisI’m excited to finally share what I’ve been building for the last year with my co-founder, Kushagra Sinha and our team: Introducing Noon. A new home for product design. We’ve raised $44M from Chemistry, First Round Capital, Scribble Ventures, SV Angel, Afore Capital, Elevation Capital, Homebrew to help us rebuild product design from the ground up. No workflow is more broken and fragmented in 2026 than the product designers’. That’s why we built Noon. The first product design tool that works entirely on your product code, so you can design how it looks and how it works. With AI at its core that works in seconds, not minutes. For the first time, you can create, iterate, build, test and ship. All in one canvas. No translations or roundtrips to the codebase and back. We are deeply grateful for the support of incredible founders and product and design leaders: Katie Dill Ian Silber Soleio Mike Davidson Ethan Eismann Michael Holzer Henry Modisett Johnnie Manzari Julie Zhuo Malthe Sigurdsson Lauryn Motamedi Scott Belsky Hemal Shah Eric Wittman Varun Anand Eric Glyman Alex Pall Romain Huet Dan Becker Daniel Scrivner Arash Ferdowsi Micah Sivitz Colin M Evans and many more. Thank you for believing in us from day one. If you’re a product designer tired of workarounds between code and canvas, Noon is for you: https://noon.design/ Comment “Get Noon” and we’ll get you on the list for early access.
-
Sanket S. shared thisContinuous CVE monitoring and alerting has landed in DeepSource. This is a big step towards helping customers who use DeepSource SCA reduce time to fix for security vulnerabilities faster.Sanket S. shared thisStudies suggest that most orgs take ~6mo on average to fix CVEs, but attackers exploit new CVEs in <5 days. DeepSource's new continuous CVE monitoring helps security teams change that. Our SCA engine now monitors vulnerability databases every hour. When a new CVE matches a dependency in your codebase, affected repos are re-scanned automatically and your team gets an email alert, so you can take action faster. Available now to all customers on DeepSource Cloud.
-
Sanket S. posted thisThe LiteLLM incident is a reminder that modern code security is no longer just about reviewing source code. The malicious payload reportedly executed through a `.pth` file on Python startup. No `import litellm` required. In other words, the compromise lived in the software delivery path, not in the application logic most teams are trained to inspect. That distinction matters. We still talk about code security as if it’s primarily a code review problem: find insecure patterns, catch vulnerable dependencies, block bad PRs. But the real trust chain is much larger: • source code • dependencies • CI/CD pipelines • publishing credentials • package registries • build artifacts • runtime access to secrets If any one of those layers is compromised, “the code looked fine” stops being a meaningful defense. AI is only going to amplify this. Teams are shipping more code, with less human scrutiny per line, across increasingly automated pipelines. That makes software trust and provenance just as important as static analysis. The next generation of code security has to answer bigger questions: • what exactly was built? • where did it come from? • who published it? • what can it access at runtime? That’s the shift the industry still underestimates.
-
Sanket S. shared thisYesterday, we shipped bring-your-own-key support for AI Review on our Enterprise Server offering. TL;DR - Run state-of-the-art AI code review and code security in your self-hosted environment, using your own keys, and with full control of how your source code is shared - Support for OpenAI, Anthropic, and Gemini models - Painless installation; just flip a switch and add your key. Shipping for a self-hosted version has always been a key design objective for us since day one of building DeepSource. Which is why, to the dismay of our infrastructure team, we don't rely on any managed services unless there's a drop-in replacement we can ship to the Enterprise Server offering. When we shipped AI Review four weeks ago, we broke this rule. We wanted to ship fast, so we could release all agentic features only on DeepSource Cloud. The traction from our customers has been incredible, and we knew we had to move quickly on bringing this to some of our largest customers who self-host. If you're looking to self-host AI Code Review for your team, reach out! If you're in SF, Jai and I will come down to your office and Collison-install DeepSource for you.
Sanket S. commented on a post
1mo
Would love your thoughts on DeepSource. We use a hybrid architecture where we use SAST as the first pass and static analysis infrastructure to give the agent more tools to explore the code. In our benchmarks runs on the OpenSSF CVE benchmark, we outperform all AI-only code security reviewers.
deepsource.com/benchmarks
-
Sanket S. liked thisSanket S. liked thisIntroducing the Polytomic Bulk Sync Activity Log. - Who turned off this sync? - Who disabled that column? - When was this table added? When syncing to your data warehouse, the larger the team the more likely these and other questions are asked. Well, wonder no more. Polytomic's newest Activity Log answers these questions. Rich timeline diffs that also come with filters to restrict your view to a particular change type or person. This concludes our support for full audit logs covering all types of syncs in Polytomic, whether to storage (data warehouses, databases, and cloud buckets) or to SaaS platforms (CRMs, marketing platforms, spreadsheets, etc)!
-
Sanket S. liked thisSanket S. liked thisTook the Strada SF team to Oracle Park last Friday! Giants vs. Cubs, fun seats, a few hours completely away from laptops and demos. We've been heads down building for a while now. It felt good to step back, get everyone together outside the office, and celebrate as a team. We didn't get the W, but the surprise 15 minute fireworks show at the end made up for it. Lights down, sky open over the bay, all of it set to music. I think about how hard we push, how much we ask of each other. Nights like this are the reason it's worth it. You can build a company with almost anyone. Doing it with people you actually want to be around is the rare part.
-
Sanket S. liked thisSanket S. liked thisLet's go! We're out of stealth with Tolmo! AI is completely redefining how software gets built. And how it's being attacked. Helping define what security means for this new world was impossible to resist. We had to go back! We’re building the AI agent fleet to secure production environments. Our security agents continuously work on a live production knowledge graph that connects data from code, cloud services, containers, databases, and third-party providers. Imagine high fidelity pictures of what you’re building, with a tracked history of changes at scale! Our team spent years on the offensive side, breaking into cloud services, kernels, or cryptographic implementations, as well as building observability and security solutions for thousands of customers in production at Datadog. Over the last few months, we’ve raised $22M, led by Accel and Y Combinator, assembled a stellar team in San Francisco and have been iterating with our first 45 design partners. It’s go time! Joining me today to announce Tolmo: my cofounders Jean-Baptiste, 📍Vladimir, and Arnaud, alongside Matt and our early team William, Jimmy, Mihail, Rania, and Tyler. We're psyched to embark on this new journey together! We're building Tolmo in the open, and we're looking for feedback. Would love to hear your thoughts about what we're building! https://lnkd.in/gFQjCBpbHello. Tolmo. $22M to build the Agent Fleet Securing Production Environments · TolmoHello. Tolmo. $22M to build the Agent Fleet Securing Production Environments · Tolmo
-
Sanket S. liked thisSanket S. liked thisI'm excited to announce that Panther has entered into an agreement to join Databricks. This is an incredible milestone and a moment I'm deeply proud of. Since 2018, we've had the opportunity to support and learn from the world’s most innovative security teams. Together, we pushed security operations forward with detection-as-code to agentic workflows across investigation and response. As AI transforms how security teams operate, it’s clear the future of security will be built on an open foundation of data and intelligence. Databricks has pioneered that foundation, and we're excited to join them in shaping the future of the security lakehouse. To our customers, investors, and every Pantherino: thank you. Your ownership, support, and hard work made this possible. Today we celebrate what we've built, and look forward to an exciting new chapter together with Databricks.
-
Sanket S. liked thisSanket S. liked thisPersonal update: after co-founding 43 almost exactly five years ago and deploying our first fund, I’ve decided to start a new chapter. I’m incredibly grateful for the experience of building 43 from the beginning alongside Dustin and Anabel, and I’ll be cheering them on as they continue building 43 from here. This experience has shaped how I think about investing, company building, and how to best support founders. Most of all, I’m grateful to the founders who let me be part of their earliest days. The honest truth is that I get so much energy from working with founders that I would do it for free. A huge thank you to all of the LPs, founders, co-investors, and friends who have supported 43 over the years. I’ll be sharing more soon on what’s next. I’ve learned so much over the past five years, and I’m excited to continue backing and supporting founders at the very beginning.
-
Sanket S. liked thisSanket S. liked thisToday is an important day for Will Reed. After more than a decade building this firm, our founder, Paige Robinson, is retiring. It's hard to overstate the impact she's had – not only on Will Reed, but on the founders, investors, leaders, and companies she's partnered with throughout her career. Paige built a firm grounded in excellence, integrity, and genuine care for people. Those values are deeply embedded in who we are and will continue to guide us moving forward. I'm honored to step into the role of Managing Partner & CEO and lead our next chapter alongside Hannah Yeargan. I’m grateful to continue building with: Jessica Bernido, Carolyn Bates Hogan, and Kelly Schulze and am thrilled to welcome Cate Fisher to the partnership. Together, they bring decades of experience helping build and scale exceptional technology companies. This transition comes at a remarkable moment for our business. Demand for leadership talent across AI-native companies, vertical SaaS, cybersecurity, fintech, and other high-growth technology sectors continues to accelerate. Founders and investors are navigating one of the most transformative periods in technology history, and we're grateful for the opportunity to help them build the teams that will define what's next. Most of all, I'm grateful to Paige – for her leadership, mentorship, and the foundation she built. We have big shoes to fill, but we're excited about what lies ahead. And thank you to our clients, candidates, and partners for trusting us to be part of your journey. The best is yet to come. 🎉 🎉 🎉
Experience & Education
-
DeepSource
View Sanket’s full experience
See their title, tenure and more.
or
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