About
I work with founders and leadership teams to design and build AI systems where human…
Articles by Mehrab
-
Something Big Is Happening. Don’t Shrink Yourself to Match It
Something Big Is Happening. Don’t Shrink Yourself to Match It
AI is accelerating fast. The real risk isn’t job loss.
3 Comments -
The Psychopath and the Language Model: Why We’re Far from AGI
The Psychopath and the Language Model: Why We’re Far from AGI
There is a lot of talk about how close we are to achieving Artificial General Intelligence (AGI). But in reality, we…
1 Comment -
Developer Skills in the AI Era — Part 2: Problem Decomposition and AI Collaboration
Developer Skills in the AI Era — Part 2: Problem Decomposition and AI Collaboration
In the previous article of this series, we delved into the crucial skill of Prompt Engineering, exploring how precise…
1 Comment -
Developer Skills in the AI Era — Part 1: Prompt Engineering
Developer Skills in the AI Era — Part 1: Prompt Engineering
In my previous article, we explored the game-changing tools revolutionizing the world of AI-assisted coding. While…
3 Comments -
AI-Assisted Coding – Tools and techniques that are changing the game
AI-Assisted Coding – Tools and techniques that are changing the game
Unless you've been living under a rock for the last year and a half, you know that AI-assisted coding is here. Many…
5 Comments
Activity
4K followers
-
Mehrab Momin shared thisI’ve been putting together this Friday’s AI Circle Houston ✨ session on computer vision, and it’s made me rethink how I explain “AI that sees.” The instinct is to treat it as one capability. It isn’t. There’s a real ladder recognizing something is there, locating it, tracking it across time, understanding what’s happening, reasoning about why, and eventually deciding to act. Most vision systems today sit somewhere in the middle of that ladder, and the language models bolted onto them can describe a scene fluently while still getting basic things wrong - like who’s holding what or which direction someone moved. If you’ve ever wondered what a camera “understands” versus what it just detects, come think it through with us this Friday at 9:30 AM at the Ion District.
-
Mehrab Momin reacted on thisThis Friday we're digging into one of AI's most underrated topics: computer vision, from raw pixels to visual intelligence. I will walk through why "seeing" isn't one skill but a ladder — recognition, detection, tracking, understanding, reasoning, action — and why a fluent, confident description of a scene is not the same as actually understanding what's happening in it. Full details here. https://lnkd.in/gj3-gr_HMehrab Momin reacted on thisAI can describe a photo in an instant. But does it actually understand what’s happening in it? This Friday at AI Circle Houston ✨, Mehrab Momin is leading a session on computer vision- not yesterday’s AI, but the unfinished problem underneath multimodal AI, robotics, and physical intelligence. We’ll trace the real arc of the field: pixels → learned features → objects → motion → actions → language → world models. Along the way: why ImageNet mattered, how video adds the dimension of time, why a system can recognize a motion without understanding why it happened, and where CLIP and vision-language models (VLMs) changed what’s possible and where they still quietly get things wrong. We’ll close on the frontier: world models and predictive architectures like JEPA, which are trying to move machines from describing what’s visible toward anticipating what happens next. Whether you’re building with AI, evaluating it, or just trying to keep up, come build a clear mental map of where visual intelligence has been and where it’s headed. Friday at 9:30 AM. Ion District. No registration needed. Bethel Kameni Joe Alapat Joey Sanchez Wade Pinder John (JR) Reale Natara Branch Dd Budiharto Ian Uriarte andre j. hermann Devina Bhojwani Layla Lumpkin Gary Martin Marisabel Escalona Whitney Adinah Zilton Salim Gheewalla Laura Neder Matthew Peña Jesse Martinez Ion District Cup of Joey Startup Circle Houston 💫 ⭕️Cybersecurity Circle⭕️ ◎⚡ Energy Circle Houston ◎⚡ Houston Storytelling Circle #AICircleHouston #AIHouston #ComputerVision #VLM
-
Mehrab Momin reacted on thisThis past week at AI Circle Houston, Tom Haymes led a sharp session on the oracle trap. Most people use AI like this: Prompt, Output, Accept, Move on That may be faster, but it does not automatically make the work better. Tom's point was that AI can quietly weaken the judgment we need to catch mistakes, challenge assumptions, and improve outcomes. The better pattern is to use AI as an augmentation tool: Verify the claim that matters most. Red-team important outputs. Review your own questions, not just AI's answers. Build deliberate friction into the workflow. The goal is not just faster work. The goal is better thinking, better judgment, and better work. Full recap here https://lnkd.in/gCcFrFa6 Stay tuned for the topic of our Next session on Friday, July 17, 2026Mehrab Momin reacted on thisLast Friday at AI Circle Houston ✨, Tom Haymes challenged one of the most common ways people use AI: Treating it like an oracle. Ask a question. Get an answer. Accept it because it sounds right. Move on. Tom called this the oracle trap. The danger is not just that AI can be wrong. The deeper issue is that frictionless AI use can disconnect us from our own work. If AI makes us faster, but we stop practicing the judgment needed to catch mistakes, question assumptions, and improve the work, then speed becomes a trap. The faster you get, the less you practice the skill that catches the expensive error. Tom framed AI as a tool for augmentation, not just automation. That means using AI to strengthen human thinking, not bypass it. A few practical habits: - Trust but verify. Do not verify every claim. Verify the one that would hurt most if it were wrong. - Red-team your work. Do not ask AI, "Are there any issues?" Ask it to take a real adversarial position. - Review your own thinking. Take the full chat log and ask AI to analyze your questions, not just its answers. Where did you stop too early? Where did you never ask why? The group discussion went deep: Bloom's Taxonomy, DIKW, RAG vs broad LLMs, local models, data privacy, the FIFA World Cup Pink Boots Problem, and how AI can accelerate convergence if everyone optimizes toward the same answer. The core takeaway: AI should not just help us produce more. Used well, it should help us think better, question better, and build better judgment. Conversation is the starting point. The real value is what people take back into their work. Mehrab Momin Joey Sanchez Joe Alapat Wade Pinder John (JR) Reale Devina Bhojwani Layla Lumpkin Dd Budiharto Marisabel Escalona Gary Martin Whitney Adinah Zilton Ian Uriarte andre j. hermann Natara Branch Matthew Peña Laura Neder Bethel Kameni Startup Circle Houston 💫 ⭕️Cybersecurity Circle⭕️ Houston Storytelling Circle ◎⚡ Energy Circle Houston ◎⚡Cup of Joey Ion District #AICircleHouston #AI #GenerativeAI #AIAugmentation #AIWorkflows #CriticalThinking #FutureOfWork #HumanAI
-
Mehrab Momin reacted on thisThis Friday at AI Circle Houston, Tom Haymes is leading a session on "Beyond a Faster Horse — Creating Quality Over Quantity with AI". Most AI conversations focus on speed: write faster, summarize faster, ship faster. This session takes the opposite view. Faster output is not the same as better work. We'll explore The Augmentation Cycle, Getting Beyond the Flawed Oracle, and how to avoid the oracle trap: treating AI's first confident answer like settled truth. The practical shift: move from Prompt -> Output -> Submit to Draft -> Question -> Red-Team -> Refine -> Decide. Full details here https://lnkd.in/g4MvT8yA
-
Mehrab Momin shared thisThis week's AI Circle Houston ✨ session is one I think a lot of teams need right now. AI is already in everyone's hands, but most people are still using it like a shortcut machine: ask once, accept the answer, move on. That can make work faster, but it does not automatically make the work better. Tom Haymes is joining us Friday for "Beyond a Faster Horse — Creating Quality Over Quantity with AI". The session will focus on using AI for true intellectual augmentation: questioning outputs, red-teaming results, challenging assumptions, refining ideas, and improving judgment before work is submitted, shipped, or acted on. Tom has been writing about this in "Making Stochastic Parrots Extinct," and the core point feels timely: AI has been handed to everyone, but the deeper challenge is agency, judgment, and knowing how to think with it well. The phrase I am holding onto is deliberate friction. If AI makes every first draft easier, we need better habits that keep us close enough to the work to catch mistakes, question assumptions, and make better decisions. If you are trying to figure out how AI changes education, software development, teams, and knowledge work, this should be a sharp conversation. Friday 9:30 AM, Ion District, 4201 Main St., Houston, TX. Just walk in.
-
Mehrab Momin reacted on thisBeen watching the World Cup and the pink boots sent me down a rabbit hole. Ended up writing a full piece on AI, convergence, and why everyone optimizing with the same models lands in the same place. Just posted it on LinkedIn. If it resonates comment or bring your disagreements Friday, this one’s worth arguing about. https://lnkd.in/gMxTDnTg
-
Mehrab Momin shared thisWatch any match at this World Cup and you'll see it before you see the football. Nike players in pink boots. Adidas players in pink boots. Puma, New Balance, Skechers. Pink. Five rival brands. Different design teams, different budgets, different sponsorship wars. All are spending millions to stand out. All independently arriving at the exact same electric fuchsia. The official explanation: a trend agency forecast the color, everyone read the same report, case closed. I didn't buy it. So I pulled the thread. What I found: trend agencies were forecasting in 2018 and 2022 too, and brands diverged both times. Stranger still, the 2026 forecast didn't even say pink. It named teal as the color of the year and listed fuchsia as an accent. Every brand ignored the forecast and made the same deviation from it, in the same direction, at the same time. Something changed between the last World Cup and this one. And I don't think it's fashion. The uncomfortable possibility I keep circling: future markets may not fail because everyone was wrong. They may fail because everyone was right in exactly the same way. I wrote the full investigation. What actually drove five companies to one color, the signal hiding in the data nobody will say out loud, and the one figure on the pitch who stood out, and why. Full piece linked in the first comment. #ArtificialIntelligence #Strategy #AIStrategy #DecisionMaking #WorldCup2026
-
Mehrab Momin reacted on thisNo AI Circle Houston session this Friday, July 3, in observance of the Independence Day long weekend. Our next AI Circle session will be Friday, July 10, 2026. Tom Haymes will lead the session: “Beyond the Faster Horse: Designing AI Workflows for True Intellectual Augmentation” Looking forward to seeing everyone then.
-
Mehrab Momin reacted on thisIf you missed Friday's AI Circle Houston session, Avleen Kaur, PhD walked us through the math behind Physical AI: why robots, drones, industrial systems, and scientific models need to handle noise, physics, limited memory, and real-time constraints. Big takeaway: Physical AI needs more than bigger models. It needs better math, better approximations, and systems that can survive the messiness of the real world. Avleen also agreed to share slides and a Jupyter notebook. Those will be posted here for anyone who wants to explore the stochastic rounding and PINN ideas more deeply. Full recap link in comments.
Mehrab Momin replied to a comment
3d
Avleen Kaur, PhD thank you for sharing your knowledge with the group.
-
Mehrab Momin reacted on thisMehrab Momin reacted on thisI’m pleased to share that I received the Ralph W. and Leona B. Gerard Graduate Student Award at this year’s Graduate Honors Banquet at UC Irvine. The award recognizes excellence in graduate student achievement, and I’m grateful to be recognized alongside so many accomplished members of the UCI community.
-
Mehrab Momin reacted on thisMehrab Momin reacted on this🌟 Ismaili Innovator Spotlight: Reza Satchu 🌟 Meet Reza Satchu — a serial entrepreneur, investor, and educator whose companies have sold for billions, but whose deepest legacy may be the thousands of founders he's shaped. Reza was born in Mombasa, Kenya, and came to Canada at age seven when his family emigrated from East Africa as part of Prime Minister Pierre Trudeau's policy to welcome Ismaili refugees. His family settled in Scarborough, where his father worked as a real estate agent and his mother as a secretary. Over two decades, he co-founded and built six operating businesses that have collectively sold for more than $3.5 billion — including StorageNow (one of Canada's largest self-storage companies), KGS-Alpha Capital Markets, L.P. (sold to BMO), SupplierMarket (sold to Ariba for $925M), and Alignvest Student Housing (Canada's largest student housing operator, sold for $1.7B). Today he is Founder and Managing Partner of Alignvest Management Corporation. But ask Reza about his greatest impact, and he'll point to his students. For over 20 years — at Harvard Business School, the University of Toronto, and NEXT Canada — he has taught aspiring founders how to calibrate risk, make decisions with imperfect information, and learn from the consequences. Collectively, his students have founded over 500 ventures, raised over $10 billion in capital, and created more than 10,000 jobs. In 2010, he founded NEXT Canada, a philanthropic entrepreneurship program now 1,200+ alumni founders strong. Notably, all of his teaching at the University of Toronto and NEXT Canada has been done for no compensation — purely as a charitable endeavor. He also spent 11 years as Vice Chairman of the SickKids Hospital Foundation and now hosts the Founder Mindset podcast with the HBS Foundry. From a seven-year-old immigrant in Scarborough to one of the most influential entrepreneurship educators of his generation — Reza Satchu embodies what it means to build, and to give back. #IsmailiInnovators #Entrepreneurship #VentureCapital #HarvardBusinessSchool #NEXTCanada #FounderMindset #Leadership #Innovation
-
Mehrab Momin liked thisMehrab Momin liked thisWhat if the next great economic shift isn't just the AI economy, but the relational economy? I've been reflecting on this question after reading recent essays by economists David Deming, Alex Imas, Kirk Doran (winner of the Adam Smith Prize), and revisiting the work of my Stanford colleague Erik Brynjolfsson. Each approaches the future from a different perspective, yet together they point toward a remarkably consistent idea. David Deming argues that humanity's greatest competitive advantage has never been intelligence alone, but our extraordinary capacity for social learning. Civilization advances because we learn from one another, through teaching, mentorship, trust, collaboration, and shared discovery. Relationships are not peripheral to innovation; they are the infrastructure through which innovation spreads. Alex Imas asks an equally profound question: What becomes scarce when intelligence becomes abundant? As AI increasingly democratizes access to knowledge, coding, writing, analysis, and even creativity, economic value shifts toward what remains uniquely human. For years, Erik Brynjolfsson has argued that AI's greatest promise is not replacing people, but augmenting human capabilities. The biggest gains will come from redesigning work so that humans and AI complement one another, allowing people to do more of what humans do best. And Kirk Doran extends this thinking even further, arguing that relational economies are essential both to sustaining long-term productivity and to understanding many of today's economic challenges. Relationships are not simply social goods—they are productive assets. Taken together, these ideas suggest something larger than a changing labor market. Perhaps we are entering a relational economy, one where trust, belonging, mentorship, collaboration, and our ability to help others learn become increasingly important drivers of innovation, resilience, and human flourishing. If that is true, then education faces a profound challenge. Across schools, universities, and workplaces, we are rightly investing enormous energy in AI literacy, helping people learn how to interact effectively with increasingly intelligent machines. That work is essential. But are we investing with equal ambition in teaching people how to relate well to other humans? Do we intentionally teach students how to listen deeply, build trust, navigate conflict, collaborate across differences, repair relationships, and cultivate belonging? This is one reason I have been exploring Relational Intelligence (RQ) as a possible next frontier of education in the age of AI. Perhaps the greatest opportunity before us is not simply to raise a generation that is AI literate. It is to raise a generation that is relationally intelligent, equipped to build the kinds of relationships that enable people, communities, organizations, and economies to thrive.
-
Mehrab Momin reacted on thisMehrab Momin reacted on thisI’m honored to join the National Board of Directors of SCORE Mentors to serve America’s largest network of volunteer business mentors. Since 1964, SCORE has helped 20 millions of entrepreneurs start, grow, and sustain successful businesses. As a resource partner of the U.S. Small Business Administration, SCORE provides free mentoring, education, and practical support that help turn ideas into enterprises and strengthen local economies. SCORE's mission is deeply personal for me. My journey from international student to business professor and venture builder, now working to shape the future of community intelligence, is a living testament to the power of #mentorship, #education, and #community. I hope to devote much of my energy to paying forward in the decades ahead. America’s 250th anniversary is a pivotal moment. Entrepreneurs have always helped write the nation’s story, and they will play an even greater role as technology reshapes how we learn, work, and live. Small businesses are more vital than ever as the backbone of our economic strength and societal wellbeing. Thank you so much to all the Board members, to Executive Director Thomas Kim and team, and to the entire SCORE community for the trust and warm welcome. I am deeply grateful to Tim Jeffcoat for the nomination and to Ricardo Casas, Jim Helfgott, Patricia Loftus, Julie Poland, Julie Christiansen, and Bryan Eckle for the generous guidance during onboarding. I am excited to begin this journey alongside fellow new board members Kelly Spillane, Rachel Wei West, and Carlos Blanco. I have been shaped by incredible mentors whose wisdom and service I truly admire - SBA Houston District Office leaders Mark Winchester and Winston Labbé and SCORE Mentors Houston leaders Willy Verbrugghe, Christian (Chris) Mahoney and Andreas Goldschmid. I am humbled to have the opportunity to serve alongside leaders I have long looked up to. I look forward to helping SCORE scale its impact, embrace technology, engage with stakeholders in the communities, and empower even more entrepreneurs across the country to shape the future together. Follow SCORE Mentors and connect with your local chapter to find a mentor, become a mentor, or support an entrepreneur in your community. 💪🇺🇸🚀 #SCOREMentors #Entrepreneurship #SmallBusiness #CommunityImpact Photo credit: University of St. Thomas (TX)
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Licenses & Certifications
-
-
Certified Secure Programmer (ECSP)
HackEDU
Patents
-
Drive-thru order facilitation system and method of use
Issued US 11,049,198
See patentA drive thru order facilitation system includes a network, having a server; a computing device; and a database; a sensor to scan an identifying feature of a vehicle; an order history platform accessible from the computing device, the order history platform having a new customer creation portal; and an existing customer profile development portal; the order history platform provides access to customer profiles; and the order history platform collects data relating to order history; the sensor is…
A drive thru order facilitation system includes a network, having a server; a computing device; and a database; a sensor to scan an identifying feature of a vehicle; an order history platform accessible from the computing device, the order history platform having a new customer creation portal; and an existing customer profile development portal; the order history platform provides access to customer profiles; and the order history platform collects data relating to order history; the sensor is to communicate with the computing device; and the computing device is to correlate the identifying feature of the vehicle with one of the customer profiles.
-
DATA SAMPLING METHOD FOR DETERMINING SALIENT VALUES IN A LARGE DATA STREAM
Issued US 9,372,603
See patentData sampling method for determining salient values in a large data stream - Techniques are disclosed for presenting users with relevant data and statistics related to the operations of an industrial system. Data obtained by a SCADA system may be stored in a real-time status database (which reflects the current process data of the industrial system) and a historian database (which archives the process data from the industrial system as it changes over time). A user may request from the SCADA…
Data sampling method for determining salient values in a large data stream - Techniques are disclosed for presenting users with relevant data and statistics related to the operations of an industrial system. Data obtained by a SCADA system may be stored in a real-time status database (which reflects the current process data of the industrial system) and a historian database (which archives the process data from the industrial system as it changes over time). A user may request from the SCADA system process data obtained during a specified period of time. The request may be subdivided into a plurality of time intervals, and process data associated with each time interval may be analyzed to determine one or more salient values that are representative of the interval. The salient values associated with each time interval then may be transmitted to the user in response to the process data request.
Languages
-
English
Native or bilingual proficiency
-
Urdu
Native or bilingual proficiency
-
Gujarati
Professional working proficiency
Recommendations received
19 people have recommended Mehrab
Join now to viewOther similar profiles
Explore more posts
-
Heather White
I keep coming back to this piece on the design space for AI coding tools. The part that stuck: AI UX isn’t one decision—it’s a stack of choices about who initiates, what context the system sees, how much autonomy it gets, and how much explanation the user needs. https://lnkd.in/g6mctx3S I tried a tiny experiment after reading it: in a sprint review, we paused on one AI feature and answered just two questions— * When should the system speak first vs. wait to be asked? * What “receipts” would make the suggestion trustworthy? That five-minute check changed the UI we shipped: fewer auto-prompts, clearer “why we suggested this” notes, calmer users. If you’ve built AI features lately, where did you land on autonomy vs explainability—and what made the choice for you? #AIUX #ProductDesign
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content