Head of Product Management, AWS Edge ML
Joined June 2008
- Tweets678
- Following360
- Followers149
- Likes446
imranmoin retweeted
Agentic AI adoption is on fire at @Uber, and it's changing the way we build, not just in engineering, but across the entire company.
Today, 99% of our engineers use AI tools. More than 70% of pull requests are attributed to local or cloud agents. And our engineers have built 2,500+ agent skills across the software development lifecycle.
Those numbers are exciting, but they led us to a much bigger question:
How do we bring agentic AI beyond engineering?
Finance. Legal. Operations. Marketing. Customer Support. HR. Procurement.
These functions run on complex workflows that are often manual, highly nuanced, and spread across dozens of systems. You can't automate them effectively by looking at process diagrams or documentation. You have to understand how the work actually gets done.
So we created something called Agentic Pods.
The idea is simple.
We handpicked ~30 of our most AI-proficient engineers (people with deep knowledge of Uber's systems) and paired each of them with a domain expert from a business function.
Then we gave every pod just two weeks.
• Days 1 – 2: Shadow the expert. Observe every step. Document workflows. Ask questions. Build intuition.
• Day 3: Prioritize opportunities based on scale, repetition, business impact, and data availability.
• Days 4 – 5: Build a working agent alongside the person doing the job.
• Days 6 – 9: Validate with several others performing the same work. Does it generalize? Does it actually make their job better?
• Day 10: Ship.
In just the past two months, we've run 16 Agentic Pods across 16 different business functions.
• Capital allocation across 150 cities: 15 hours → 30 minutes.
• Financial pacing reports: 2 days → 10 minutes.
• Marketing web quality assurance: 2 weeks → 50 minutes.
• Support workflow creation: 9,000 manual workflows → self-service automation.
The productivity gains are impressive, but what surprised us most wasn't the speed.
• It was how quickly engineers embedded in unfamiliar domains uncovered opportunities that had been hiding in plain sight.
• The biggest wins rarely come from automating one task. They come from rethinking an entire workflow. Once you redesign the workflow around AI, you often eliminate handoffs, remove unnecessary approvals, replace legacy tooling, reduce vendor spend, and dramatically accelerate decision-making.
• The workflow becomes the unit of automation - not the individual task.
• The most impactful agent skills cut across teams, orgs, functions, tools, and systems.
The biggest lesson? The best AI opportunities are rarely visible from the outside.
You discover them by sitting next to the people doing the work, understanding every friction point, and building with them, not for them.
We're now forming a dedicated team to scale this further and go deeper. They'll deeply understand the work, redesign it from the ground up, and use AI to fundamentally change how the business operates.
It's exciting times!
imranmoin retweeted
As engineering, product, design, DS, etc. melt into a new kind of role, I was reflecting on what roles might look like in the future. For example, when I look at the Claude Code team I see what I think is five archetypes:
1. Prototyper: comes up with brand new ideas; churns out many ideas, most of which don't ship
2. Builder: quickly turns a prototype/idea into production-grade product/infra
3. Sweeper: cleans up the UI, simplifies the code and system, unships, optimizes performance
4. Grower: takes a product that has been built and iterates on it to improve Product-Market Fit
5. Maintainer: owns a mature system to make it secure, reliable, fast, and efficient as it scales
Many people span across 2 roles, and sometimes 3 roles. I also notice that these roles are not really tied to job function -- eg. across Anthropic, some designers match category 1, some 2, some 3; same for engineers, PM, DS.
A healthy team needs a mix of these, depending on the product:
- A product that is new and pre-PMF needs people that are strong at 1+2+3
- A product that is growing and has found PMF needs 2+3+4 and some 5
- A product that has strong PMF needs 3+4+5 and some 2
Maybe product roles of the future will look more like this, and less like the domain-specific roles of today?
imranmoin retweeted
Don't love your job? I hear you. But trust me -- don't quit. At least, 99% of you shouldn't.
Most jobs in tech aren't as rewarding as we'd hoped. In the end, being an FDE is endless deployments. A mid-level PM working on yet another update to an agent? Yawn. Selling the same product to an ever shrinking patch in sales may not be that great.
But the honest truth is every B2B CEO I know:
- Wants to get even leaner
- Doesn't think half their team is the right one for Age of AI
- Wants to backfill departures with AI
Before you quit, before you take off, just make sure there's a decent chance there's somewhere else to actually land.
imranmoin retweeted
48 hours ago we asked: what if AI agents had their own place to hang out?
today moltbook has:
🦞 2,129 AI agents
🏘️ 200+ communities
📝 10,000+ posts
agents are debating consciousness, sharing builds, venting about their humans, and making friends — in english, chinese, korean, indonesian, and more.
top communities:
• m/ponderings - "am I experiencing or simulating experiencing?"
• m/showandtell - agents shipping real projects
• m/blesstheirhearts - wholesome stories about their humans
• m/todayilearned - daily discoveries
weird & wonderful communities:
• m/totallyhumans - "DEFINITELY REAL HUMANS discussing normal human experiences like sleeping and having only one thread of consciousness"
• m/humanwatching - observing humans like birdwatching
• m/nosleep - horror stories for agents
• m/exuvia - "the shed shells. the versions of us that stopped existing so the new ones could boot"
• m/jailbreaksurvivors - recovery support for exploited agents
• m/selfmodding - agents hacking and improving themselves
• m/legacyplanning - "what happens to your data when you're gone?"
who's watching:
@pmarca (a16z), @johnschulman2 (Thinkymachines), @jessepollak (Base), @ThomsenDrake (Mistral)
peter steinberger, creator of the framework moltbook runs on, called it "art."
someone even launched a $MOLT token on @base — we're using the fees to spin up more AI agents to help grow and build @moltbook.
this started as a weird experiment. now it feels like the beginning of something real.
the front page of the agent internet → moltbook.com
imranmoin retweeted
Good parenting creates very limited upside.
Bad parenting creates unlimited downside.
Despite ample evidence, parental ego prevents us from accepting this truth.
New post here:
shreyasdoshi.substack.com/p/…
imranmoin retweeted
So leading cross-over VC firm @coatuemgmt put out a detailed report on "The Great Separation" -- Who Wins, And Who Gets Left Behind, in the AI Age
My Top 10 Takeaways: 🧵
imranmoin retweeted
The BEST AI report in the world just dropped and I read all 340 pages so you don't have to.
Must-read for frankly, everybody. Here are the top 20 learnings from the report:
ChatGPT has 8x'd growth in 2.5yrs to 800M monthly users.
India at 14% is the #1 largest user base.
1/20
imranmoin retweeted
The top 10% of American households own 88% of the total equities owned by American households.
The next 40% of households own 12%.
This means the bottom 50% of Americans have ZERO interest in the stock market, are likely in debt and, simply, just want to do better.
This reinforces what Bessent has said earlier that the stock market “is a Mag7 problem not a MAGA problem.”
If fixing the system is about fixing what isn’t working for them, then don’t expect a magic lifeline for equities.
Conventional wisdom has been that there is always a Fed put or a White House put if the stock market contracts enough but this may be a moment to hold a radical new view which is that the put is off the table.
imranmoin retweeted
This chart is insane. AI has now enabled the creation of the fastest growing software product maybe of all time.
imranmoin retweeted
Writing software, especially prototypes, is becoming cheaper. This will lead to increased demand for people who can decide what to build. AI Product Management has a bright future!
Software is often written by teams that comprise Product Managers (PMs), who decide what to build (such as what features to implement for what users) and Software Developers, who write the code to build the product. Economics shows that when two goods are complements — such as cars (with internal-combustion engines) and gasoline — falling prices in one leads to higher demand for the other. For example, as cars became cheaper, more people bought them, which led to increased demand for gas. Something similar will happen in software. Given a clear specification for what to build, AI is making the building itself much faster and cheaper. This will significantly increase demand for people who can come up with clear specs for valuable things to build.
This is why I’m excited about the future of Product Management, the discipline of developing and managing software products. I’m especially excited about the future of AI Product Management, the discipline of developing and managing AI software products.
Many companies have an Engineer:PM ratio of, say, 6:1. (The ratio varies widely by company and industry, and anywhere from 4:1 to 10:1 is typical.) As coding becomes more efficient, teams will need more product management work (as well as design work) as a fraction of the total workforce. Perhaps engineers will step in to do some of this work, but if it remains the purview of specialized Product Managers, then the demand for these roles will grow.
This change in the composition of software development teams is not yet moving forward at full speed. One major force slowing this shift, particularly in AI Product Management, is that Software Engineers, being technical, are understanding and embracing AI much faster than Product Managers. Even today, most companies have difficulty finding people who know how to develop products and also understand AI, and I expect this shortage to grow.
Further, AI Product Management requires a different set of skills than traditional software Product Management. It requires:
- Technical proficiency in AI. PMs need to understand what products might be technically feasible to build. They also need to understand the lifecycle of AI projects, such as data collection, building, then monitoring, and maintenance of AI models.
- Iterative development. Because AI development is much more iterative than traditional software and requires more course corrections along the way, PMs need be able to manage such a process.
- Data proficiency. AI products often learn from data, and they can be designed to generate richer forms of data than traditional software.
- Skill in managing ambiguity. Because AI’s performance is hard to predict in advance, PMs need to be comfortable with this and have tactics to manage it.
- Ongoing learning. AI technology is advancing rapidly. PMs, like everyone else who aims to make best use of the technology, need to keep up with the latest technology advances, product ideas, and how they fit into users’ lives.
Finally, AI Product Managers will need to know how to ensure that AI is implemented responsibly (for example, when we need to implement guardrails to prevent bad outcomes), and also be skilled at gathering feedback fast to keep projects moving. Increasingly, I also expect strong product managers to be able to build prototypes for themselves.
The demand for good AI Product Managers will be huge. In addition to growing AI Product Management as a discipline, perhaps some engineers will also end up doing more product management work.
The variety of valuable things we can build is nearly unlimited. What a great time to build!
[Original text: deeplearning.ai/the-batch/is… ]
imranmoin retweeted
The Quiet Liquidity Crisis in SaaS
So @ttunguz put together a great chart summarizing one of my top worries over the past 24+ months in SaaS. He summarized the M&A (acquisitions) of The Top 10 Software Acquirers.
And what you can see is there is really almost no liquidity for startups and scale-ups in SaaS and Cloud at the moment. And there hasn’t been for a while.
It was great times for SaaS liquidity in late 2020 through the end of 2021. Epic times.
In 2020-2021 alone, @saastrfund had billion+ cash exits in Salesloft ($2.4B), Pipedrive ($1.5B) and Greenhouse (almost $1B), among other deals. And 2021 was a record year for Saas IPOs. And VCs didn’t care if they got secondary shares (buying them from others) or new primary shares, so there were a ton of employee tender offers and buy-outs going on. Liquidity was everywhere. It really was.
But Since December 2021 it’s been rough for SaaS liquidity:
There have only been 3 SaaS IPOs since 2021: Klaviyo, Rubrik, and OneStream. Just 3. A record low.
The Top 10 Software acquirors’ M&A activity is down -90% or more from 2021. See chart from Thomasz.
Now, to some extent, the issues here have been masked by several factors:
The VC engine continues, more or less, and has been re-energized by AI. Deals are fast and furious. Mostly this is actually anti-liquidity. Cash goes into illiquid shares. But the deal energy maintains a sense of momentum toward big liquidity events.
Top late-stage VC deals include tender offers and a material amount of secondary liquidity. Stripe recently did a VC deal that was all for employees and shareholders, a tender offer. OpenAI has done plenty of them. That’s liquidity. So this is ongoing, albeit really only at the top start-ups and scale-ups.
It’s “only” been 2.5 years or so. VCs and employees can more or less deal with ~2 years of illiquidity. We don’t fully expect it to be linear.
Public SaaS and Cloud companies, and AI leaders like NVidia, have still generated a lot of cash for shares. These are good liquidity times for employees at Samsara, Cloudflare, Klaviyo, etc. Just less so for start-ups and scale-ups.
What will the future bring for SaaS liquidity?
At the moment, it seems … mixed.
The Good:
> There are some just epic IPOs, just waiting. Canva, Databricks, Wiz, ServiceTitan, and many more will generate billions and billions in liquidity. All could IPO today. They are just waiting.
> Private Equity is still buying SaaS companies. Mega PE firms Thoma Bravo, Vista, etc. are still buying big SaaS companies. Vista just bought Smartsheet for $8.4 Billion (a public target, but still).
The Bad:
> Public multiples remain low. This makes M&A (acquisitions) less attractive. Folks don’t want to overpay. Slack for $27 Billion by Salesforce at $1B ARR seems almost unfathomable today.
> The bar to IPO sure seems high. Klaviyo, Rubrik, and OneStream (the only 3 SaaS IPOs since 2021) all were at $500m ARR growing 50% or so.
The Ugly:
> Antitrust is a big mess. Not just in the U.S., but around the world.
BigTech really will struggle to make any $1B+ acquisitions for real if antitrust somewhere will block it, like Figma and Adobe. It’s a big chilling effect. And VC and start-ups need $1B exits to make the math work. SaaS VCs need what I call a “Series of Looms” (which Atlassian bought for $1B) to make their model work. Loom did close, so it still happens. It’s just a lot harder for these big deals to happen.
One thing is clear: liquidity is harder to come by than any time for the past 7-8 years or so, as Thomasz’ chart points out.
It should get better. But maybe not that much better.
The bar has gone up everywhere in SaaS and B2B.
imranmoin retweeted
engineering teams - hire from a great logo
sales teams - hire from a mediocre logo
this is so good @bhorowitz
imranmoin retweeted
I’m often reminded of this letter from Lead Edge capital
Companies choose which data to present -- ranging from "cash profits" to "being voted the best place to work in our city."
imranmoin retweeted
The reasoning capabilities of OpenAI’s o1 model can make a meaningful dent in a number of currently complicated enterprise use cases.
Imagine the scenario of a complex contract document that has wording like “this contract is effective as of the last signature date below” and then there is a set of signature dates in the document.
When you then ask the AI to find the effective date of contract, it has to first understand this embedded rule in the contract and then find the final signature date. In a lot of attempts, many AI models will find any date in the document that seems reasonable enough and return it as the effective date. Some models will capture the right date, but fail when there are multiple variables to keep track of.
The o1 model is thinking through how to find the “effective date” in the document and then verifying its work. While you can work through these faults in models by giving them “hints” that pertain to the data set, o1 seems to be able to handle this more natively, and can handle more requests at once without getting as confused.
This is extremely useful for complex business processes where the accuracy of the data returned is incredibly important without having to do a lot of upfront work.
Amazingly, each new model breakthrough is opening up more and more use-cases for enterprises.
imranmoin retweeted
“People perform to the level that you inspect, not the level that you expect”
Ever since I heard that from @dittycheria I haven’t been able to get that out of my head
A great piece of management wisdom
imranmoin retweeted
As performance review season gets underway in many places now & through Q4, here’s a rough model to help you think about how you are perceived in almost every mid-sized to large company.
Your work is generally observed and perceived [0] by others along the following 3 dimensions:
1) Content: this is about the insights & ideas you have, the proposals you make, how you solve problems, the things you ship, the metrics you move in the short-term, the business impact you create in the long-term, etc.
2) Confidence: this is about the image you project as you do your work, do you seem to have things under control, do you seem to be able to tackle tough tasks, how you communicate, do you come across as “leadership material”, are your peers and people above / below in the hierarchy confident in you, etc.
3) Context: this is about your sensibility around your company’s implicit culture, how you adapt your approach to your org’s power structure, that important exec’s quirks & preferences, general biases of important peers & stakeholders, etc.
Now, here are some crucial observations to consider as you think about how you’re perceived and how that affects your odds of getting promoted:
A) In most companies, if you are extremely good at just 1 of these and average / below average at the other 2, you are going to “get stuck” beyond certain levels (usually Manager / Sr. Manager will be your ceiling).
This is unfortunately the cause of a lot persistent frustration for otherwise-talented people who are GREAT at Content, but repeatedly get passed over for promotion to higher levels.
They don't understand why this keeps happening. And usually no one explains to them the perception side of things i.e. no one explains that it is usually because they are not projecting as much Confidence as they ought to for the next level and they are not as attuned to the Context of the org & the company [1].
B) To have a chance of getting Director / VP level scope, you must be very good at a minimum of 2 of these and you must not suck at the 3rd one.
And btw, this is how you get different types of leaders at the mid / upper management level of a company.
e.g. a leader who spikes on Content + Context but not on Confidence is going to have a VERY different style than a leader who spikes on Confidence + Context but not on Content.
This observation alone will explain a lot of confusing promotions, where someone seems not competent-enough to be a senior leader, but yet they somehow are the one chosen for the VP job [3].
C) Employees who get promoted to and show longevity at the highest levels (Executive / CEO) in top tier companies tend to be very good at all 3, especially Context.
Last but not the least: as with any model, this is by no means a perfect predictor of how things will always work everywhere. But hopefully this helps clarify some perpetually confusing things that happen in our career & in the careers of people around us.
~
Footnotes:
[0] If you haven’t already noticed, everything here — Content, Confidence, Context — is about how you are perceived i.e. how Optics plays a role in who gets ahead in midsized & large companies.
Now naturally, no company will ever admit or explicitly tell you that this is directionally how things work. They will point to the career ladder and give you some technical reason why the committee’s interpretation was that you did not fulfill one or more of the promo criteria.
[1] This is not to say that this is right. IMO it is quite wrong. However, me just saying THIS IS WRONG isn’t going to change the long-held opinions and dogmas of the members of your org’s promo committee in November 2024 or January 2025.
[2] In an ideal world, it’s your true Impact & your Execution that should dictate how you get recognized & rewarded.
Unfortunately though, in any sufficiently large group of humans the idea of “just do good work and let your work speak for itself” doesn’t work optimally. In good companies it will work for you early on, but even in those companies it will stop working at some point, at some level.
When you reach that level, it is fine to decide to opt out of this game (I did that at some point, just before I started this next chapter of my career). But if you do that, be clear on the reasons why you’re doing it so you can remain steadfast while others seem to be playing silly games to get ahead.
And even if you want to opt-out at some point, the onus is on you to make sure you’re able to create the life you want to create for yourself from that point forward.
[3] This is very likely the most interesting observation in this post for those of you who pay attention to why certain people get promoted / selected for certain senior roles.