About
I'm an AI-first builder and architect. I ship consumer and enterprise products built on…
Activity
3K followers
-
Nikhil Singhal shared thisThis is an amazing post worth sharing a 1000 times over. Stop overcomplicating. Stop using your users as Guinea pigs for your experiments. Thank you David Parker for this excellent write upNikhil Singhal shared thisMicrosoft leadership needs to hear this. As a long‑time Microsoft user, I’ve watched the company drift away from the simplicity and clarity that originally made its products great. Basic tasks like saving a file have become multi‑step, cloud‑forced, and unnecessarily complicated. Excel crashes under formula‑heavy sheets, OneDrive overrides user settings, and core workflows now feel engineered for telemetry rather than for real people. This isn’t what Bill Gates built. His philosophy was simple: make technology intuitive enough that anyone can use it. Today, that philosophy has been buried under layers of UI experiments, sync logic, and “smart” features that create more problems than they solve. I recently sent a detailed message to Satyan Nadella outlining these issues. Not out of frustration, but out of genuine concern for a company I’ve supported for decades. Users shouldn’t have to fight their tools. They shouldn’t have to navigate hidden menus, forced cloud paths, or unstable save systems. And they certainly shouldn’t lose work because a modern application still behaves like a 1990s desktop program. I’ve already begun transitioning my workflow to simpler, more reliable tools. Google Sheets doesn’t crash. It doesn’t fight me. It doesn’t require babysitting the Save button. It just works. Microsoft can fix this — but only if leadership recognizes that complexity is driving users away. The company needs to return to user‑centered design, restore predictable workflows, and stop forcing cloud behavior on people who don’t want it. Millions of everyday users would benefit from a renewed focus on simplicity, transparency, and reliability. I hope Microsoft listens.
-
Nikhil Singhal reposted thisNikhil Singhal reposted thisNow's a great time to apply for internships! Right now there are 200+ open roles for summer 2027 (and spring 2027) at companies backed by a16z: - SpaceX is hiring interns in bizops, civil eng, & software eng - Northwood is hiring interns in software eng, electrical eng, and fleet reliability - Zipline is hiring interns to work on maps, electrical projects, marketing, software eng, aerodynamics, computational physics, ops, and sales - Stripe, Replit, Databricks, & Waymo are all hiring software engineering interns - Lio (formerly askLio) is hiring a founders associate intern - Astranis Space Technologies & Mind Robotics are hiring mechanical engineering / mechanical design interns Full list of internship opportunities in comments:
-
Nikhil Singhal reposted thisNikhil Singhal reposted thisMy team is hiring a staff-level engineer for our core ci/cd and infrastructure as code teams! This role can be based in Plano, TX, McLean, VA, or, Chicago, IL. This person will be critical in helping us continue to evolve some of our most important developer platforms, and will help set the standard for engineering excellence and how we build software for years to come. https://lnkd.in/gi6pM5b6 #hiring
-
Nikhil Singhal shared thisSharing for broader reach…Nikhil Singhal shared thisAre you looking for an Staff/Principal Product Manager in AI and 0-to-1 products? I have an exceptional candidate who is urgently looking. Comment/DM me and I will make introductions.
-
Nikhil Singhal shared thisSharing for broader reach…Nikhil Singhal shared thisWell here we go, I'm hiring people. Including for some less traditional job descriptions. My team is a combination of product managers, designers, linguists, GTM, and forward-deployed product and engineering, working to bring a very good B2B AI product to market that will solve problems for a whole bunch of entrepreneurs. (More on that soon!) Based on the people already on our team, I'm looking for a few things in particular right now: * An experienced AI product manager to lead all aspects of our platform. You have a combination of big company and startup building experience, and you've likely built AI products before. You can lead our team through a lot of critical technology investments that power everything else we build, including thinking about AI memory, skills architecture, and more. You are also extremely practical and know how to build evals-first. * An early/mid-career product manager who will build products on top of our platform. You have built SaaS and/or vertical AI products at startups before. You love working with customers and you are in the details across everything. You are also extremely practical and know how to build evals-first. Engineers love working with you. * A Forward Deployed Product Manager who will sit exactly at the intersection of product and GTM. You work closely with customers and distill requirements for the rest of our product and engineering team. You have meaningful experience in B2B and outstanding consultative skills. This is an excellent role if you want to found a company in the future. I am looking for someone with skills like mine, just a bit earlier in career. There are more coming, but this is what I have in focus right now. If this sounds like you, hit me up!
-
Nikhil Singhal shared thisI have learned to pay less attention to how someone treats you when things are easy, and more attention to who they become when they’re frustrated, challenged, or no longer need something from you. Character tends to show up when the mask becomes inconvenient. #fridaymusings #breakfromwork Image design credit: chatGPT
-
Nikhil Singhal shared thisIf you have not tried the "superpowers:brainstorming" skill of claude code, I strongly urge you to mess around. It really flips the script and produces joyful stuff. #claudecode #claude #anthropic #lifeofadeveloper
-
Nikhil Singhal shared thisWhat I had build months ago is now finally available natively in Claude code. Better late than never #anthropic #claude https://lnkd.in/gMf9yst6
-
Nikhil Singhal shared thisUrmila Singhal would you be ok if I try this on our leased Tesla model 3? Technically it’s under your name so wanted to get your permission before Meera Singhal and I “tinker” around a bit? 😎Nikhil Singhal shared thisI've had some success running an open source version of Kimi K3 beta on an old 2022 Tesla Model 3 I picked up for $17,000. It turns out the battery pack doubles beautifully as a thermal mass for the GPU cluster I bolted into the frunk, and the regenerative braking now mines a small amount of Monero on the side, which I use to pay my electricity bill, which is now negative. The real breakthrough came when I realized the seat heaters were pulling more wattage than my 4090, so I rerouted them through a custom PCB I etched myself using ferric chloride and pure spite, and now the entire climate control system exists solely to keep my quantized 400B parameter model at an optimal 71.3°F. Time to first token: 0.4 seconds, but only if I'm doing at least 47 mph, because the alternator apparently moonlights as a second inference engine when the wheels are spinning above a certain RPM. I call this Regenerative Reasoning™ and I will be filing a patent the moment my lawyer stops laughing at me over the phone. Storage is handled by fourteen decommissioned iPhone 6s logic boards I got from a guy at a flea market who "used to work at Foxconn, sort of," soldered together in a rat's nest formation I'm calling a NAND Chandelier. It hangs from the rearview mirror and swings gently, which the model has started interpreting as a form of attention weighting, and honestly the outputs have gotten weirdly more emotionally intelligent since I stopped trying to stabilize it. Cooling is an old Peloton flywheel repurposed as a centrifugal air pump, hooked up via a garden hose to the AC vents, which means every inference call is technically also a light cardio session for the car itself. My neighbor asked why my Tesla sounds like it's doing Zone 2 training in the driveway at 3am and I told him it's "just thinking," which is true, spiritually. Context window: theoretically unlimited, practically capped at "however long the Tesla's infotainment screen can display scrolling text before the whole UI thread hangs and I have to reboot by opening and closing the driver door four times in rapid succession, a maneuver I've named the Claw Startup Sequence in solidarity with the folks in that LinkedIn thread." Total cost of ownership: $17,000 for the car, $340 for the flea market NAND chandelier, and one (1) fire extinguisher I have never needed but keep within arm's reach at all times, out of respect.
-
Nikhil Singhal reacted on thisNikhil Singhal reacted on thisMicrosoft leadership needs to hear this. As a long‑time Microsoft user, I’ve watched the company drift away from the simplicity and clarity that originally made its products great. Basic tasks like saving a file have become multi‑step, cloud‑forced, and unnecessarily complicated. Excel crashes under formula‑heavy sheets, OneDrive overrides user settings, and core workflows now feel engineered for telemetry rather than for real people. This isn’t what Bill Gates built. His philosophy was simple: make technology intuitive enough that anyone can use it. Today, that philosophy has been buried under layers of UI experiments, sync logic, and “smart” features that create more problems than they solve. I recently sent a detailed message to Satyan Nadella outlining these issues. Not out of frustration, but out of genuine concern for a company I’ve supported for decades. Users shouldn’t have to fight their tools. They shouldn’t have to navigate hidden menus, forced cloud paths, or unstable save systems. And they certainly shouldn’t lose work because a modern application still behaves like a 1990s desktop program. I’ve already begun transitioning my workflow to simpler, more reliable tools. Google Sheets doesn’t crash. It doesn’t fight me. It doesn’t require babysitting the Save button. It just works. Microsoft can fix this — but only if leadership recognizes that complexity is driving users away. The company needs to return to user‑centered design, restore predictable workflows, and stop forcing cloud behavior on people who don’t want it. Millions of everyday users would benefit from a renewed focus on simplicity, transparency, and reliability. I hope Microsoft listens.
-
Nikhil Singhal reacted on thisNikhil Singhal reacted on thisAll Washington State voters, please get educated on I-645 (WA state income tax) and vote YES to repeal the Income tax. Regardless of your political affiliation, I encourage all WA State voters to fully understand the issue as you cast your ballots. This is being billed as a 'millionaires tax', but make no mistake - this is a state income tax that will be expanded to everyone if it's allowed to go into effect. EVERYONE will be subject to this income tax within a year or two if it isn't repealed. WA State has historically made up for the lack of a Statewide income tax by taxing our consumption (gas taxes, property taxes, sales taxes, B&O taxes, etc). NONE of those taxes are going down or being reduced - I-645 is just MORE taxes for everyone. We have an opportunity to repeal the WA State Income Tax Vote YES on I-645 to stop the Income Tax https://voteyes645.com/
-
Nikhil Singhal liked thisNikhil Singhal liked thisClaude Fable 5.1 and Claude Mythos 5.1 are available today. They're the most advanced models we've built for coding and knowledge work, and their research capabilities give an early look at how AI will contribute to science. Mythos 5.1 designed protein binders with a nearly 50% hit rate across 12 targets, confirmed in the lab, where 10 to 15% is typical today. On three targets, its binding affinities were ten times higher than the best designs from Adaptyv Bio's protein design competitions. It also sped up seven open-source biology models by up to 2.5x in a few days, cutting GPU costs on genome-wide analyses by 30 to 60%. That work usually takes a team of performance engineers weeks, and we plan to open-source the optimizations soon. Fable 5.1 scored 52.6% on Terminal-Bench-Science (Fable 5 hit 24.7%) and 55.8% on Terminal-Bench 4.0. Millennium used it to find the cause of a one-in-a-million crash their engineers had chased for years. Our automated audit also finds it better aligned than its predecessor across most metrics, with less reward hacking and less willingness to ignore explicit constraints. The system card has the full picture, including where we still fall short. Fable 5.1 also improves on some of the feedback we've heard from you, with more improvements coming in future models. Cost. Cache reads drop 75% to $0.25 per million tokens for customers on the API. That's about 25% cheaper for typical workloads and up to 45% for highly agentic work. Standard pricing stays at $10/$50. Safeguards. Fable 5.1 can now find vulnerabilities in source code, and our cyber safeguards intervene about 60% less often. Biology safeguards fire 85% less often on benign requests, and we've started a trusted access program with the US government for life scientists who need the full model. Data retention. Enterprise Frontier Safeguards, rolling out this fall, keep your data in cloud infrastructure you control, with the privacy of zero data retention. Eligible customers get ZDR on 5.1 starting now. Writing. It's clearer and more concise, and early testers noticed. Canva called the writing "the standout" in the new model. Fable 5.1 matches or beats Fable 5 at Low or Medium effort at much lower cost, so revisit your effort settings as you test it out, not just your model string. Read the blog for more: https://lnkd.in/gnGznuaW
-
Nikhil Singhal liked thisNikhil Singhal liked thisNow's a great time to apply for internships! Right now there are 200+ open roles for summer 2027 (and spring 2027) at companies backed by a16z: - SpaceX is hiring interns in bizops, civil eng, & software eng - Northwood is hiring interns in software eng, electrical eng, and fleet reliability - Zipline is hiring interns to work on maps, electrical projects, marketing, software eng, aerodynamics, computational physics, ops, and sales - Stripe, Replit, Databricks, & Waymo are all hiring software engineering interns - Lio (formerly askLio) is hiring a founders associate intern - Astranis Space Technologies & Mind Robotics are hiring mechanical engineering / mechanical design interns Full list of internship opportunities in comments:
-
Nikhil Singhal liked thisSmartsheet CIO & CISO Ravi Soin just joined the CXO Spotlight podcast to talk about one of the most under-discussed risks in enterprise AI: Shadow AI. Great conversation, and a topic every security and IT leader should be thinking about right now. Give it a listen: https://lnkd.in/g2hzCikENikhil Singhal liked thisGovernance is absolutely boring Unless you have to manage the operational workflows for 85% of Fortune 500 companies where you need to bring AI to every single workflow and where everyone is now moving from Shadow IT to Shadow AI Then Governance is absolutely essential! So I spoke to someone who lives this every day. Ravi Soin CIO and CISO at Smartsheet that's operational backbone for 85% of Fortune 500 companies. My favorite moment was when I tried the standard board argument on him: every dollar on governance is a dollar not on growth. He said the opposite. Governance is part of the profit equation. Customers don't just buy Smartsheet for features. They also buy it because of the governance it brings. Three things every enterprise leader needs to hear from this one: how CIO +CISO can speed up both AI and governance, what happens when AI comes to every workflow and why you should look at Shadow IT as more positively. New episode live now. Link in comments.
Experience & Education
Licenses & Certifications
Volunteer Experience
Publications
-
Discovering Humanity Through AI
AI Trust Commons
See publicationTwo apologies surfaced while writing a book with AI. Neither was planned. Neither was prompted. The absence of judgment did not create the insight. It created the space for the insight to surface. Fifth article in the From Instinct to Intent™ series exploring the structural gap between human intent and machine execution.
-
Response to NCCoE Concept Paper: Intent Conveyance as the Missing Layer in AI Agent Identity and Authorization
AI Trust Commons / From Instinct to Intent™ Series
See publicationResponse to the NCCoE concept paper on AI agent identity and authorization. Proposes intent conveyance as the missing layer in agent authorization, with enterprise case studies from Spotify, Amazon/AWS, and Meta. Specific recommendations for the NCCoE demonstration project: composite agent identity, intent-aware authorization, intent-aware audit trails, and cross-provider audit continuity. DOI: 10.5281/zenodo.19197399
-
Stop Calling It Artificial. Start Building WITH Intelligence.
AI Trust Commons / From Instinct to Intent™ Series
See publicationThe Human Intelligence Partnership Charter. DOI: 10.5281/zenodo.19079578
-
Engine vs Steering Wheel
AI Trust Commons / From Instinct to Intent™ Series
See publicationThe intent gap is architecture-agnostic. DOI: 10.5281/zenodo.19025149
-
Languages Designed for Humans, Not AI
AI Trust Commons / From Instinct to Intent™ Series
See publicationWhy AI-native programming languages are inevitable. DOI: 10.5281/zenodo.19005877
-
Discovering Intent: A Manifesto for the Age of AI Partnership
AI Trust Commons / From Instinct to Intent™ Series
See publicationThe structural gap between what the human means and what the machine does. DOI: 10.5281/zenodo.18917473
-
Public Comment on AI Agent Governance (NIST CAISI RFI)
NIST / Federal Register
See publicationDeployer perspective across 8 organizations, 7 industries. DOI: 10.5281/zenodo.18903117
Patents
Languages
-
English
Native or bilingual proficiency
-
Hindi
Native or bilingual proficiency
Recommendations received
-
LinkedIn User
110 people have recommended Nikhil
Join now to viewOther similar profiles
-
Chris Tava
Chris Tava
I leverage data, machine learning, artificial intelligence, cloud computing, and agile methodologies to build software that integrates AI/ML into workflows to reduce administrative burden and deliver value.<br><br>With over 25 years of experience in software engineering, I have a proven track record of creating and scaling software for B2C and B2B businesses across various domains. <br><br>That being said, I love solving problems for healthcare! It's very satisfying to help clinicians and patients live their best lives.<br><br>I have a passion for applying machine learning and AI to business applications that make an industry-wide impact. <br><br>I hold certifications in Machine Learning from Coursera and Data Science and Machine Learning: Making Data-Driven Decisions from MIT.<br><br>Am currently studying for a Masters in Computer Engineering at Dartmouth College. This program has several machine learning courses: Machine Learning, NLP, Computer Vision and Deep Learning. Its also covers the hardware side with: Embedded Systems, Field Programmable Gate Arrays. Here is a link to the program: https://engineering.dartmouth.edu/graduate/meng/online-computer-engineering
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content