Free with a sign-in.
The best 10 min of 32
Bringing Continual Learning to Every Company: Trajectory’s Ronak Malde
Ronak Malde says continual learning will let companies turn real-world agent experience into models that improve with use.
Training Data · Pat Grady & Lauren Reeder
- Ronak Malde is cofounder of Trajectory.
The best minutes
- 13:05Single rollouts can train agents without replay environmentsPro pick 3 min
- 20:11Continual learning keeps humans in the loopPro pick 4 min
- 10:27Continual learning starts with traces, replay and training2 min
Every moment plays free. PRO marks a write-up that's in Pro.
Also worth watching
- 4:40Harvey raised its legal benchmark score from 4% to 18%4 min
- 29:04Malde sees specialization lifting AI adoption from 1% to 99%2 min
Skip the model-architecture overview, the platform and Rogo discussion, and the extended team, product-builder, and hiring sections.
Summary · 3 min read
Ronak Malde is a cofounder of Trajectory, which aims to make model training, evaluation and deployment accessible to companies. He previously worked at Windsurf, where he says his team trained a model using production edits and retries.
Malde argues that today’s agents are pseudo-continual learners: they can perform tasks, but their models are updated in offline batches rather than learning from use. He says Trajectory aims to help companies turn production traces and their own definitions of quality into model improvements, using approaches such as group-size-one GRPO and a meta-grader to reduce environment-building and labeling work. He points to Harvey, where a model trained with Trajectory moved from 4% to 18% on its legal benchmark at six times lower cost than frontier models. He distinguishes continual learning, which keeps humans in the loop, from recursive self-improvement, and envisions product builders steering products that adapt over time.
A Pro write-up, open to you here. Try Pro free for 7 days
A Pro write-up. The rest of the summary and the minute by minute are in Real Alpha Pro; the episode, its moments and the transcript stay free.
Read it with Pro, 7 days freeMinute by minute
- 0:00Training should close the loop
Malde says Trajectory’s product philosophy draws on Apple’s combination of the internet and the phone. He wants to put training, evaluation and deployment together, then send improved models back into production to gather new traces.
- 1:00Agents still learn in batches
Malde defines continual learning as learning online, like the human brain. He compares current agents to paper road maps that require offline updates, while his goal is closer to Google Maps, which learns from routes and traffic as people travel.
- 3:08Model research needs a real-world bridge
Malde says model architectures are already changing, including through fast and slow memory and skip connections. He sees Trajectory’s role as connecting this research to company feedback and translating each company’s idea of good performance into model improvements.
- 4:30Training models was costly and slow
Malde says training a frontier model at Windsurf required hundreds of millions of dollars in R&D compute, a large research team and six to nine months. He saw production edits and retries as a repeatable source of feedback; he says Harvey used Trajectory to train a model six times cheaper than frontier models and raise its result on its legal benchmark from 4% to 18%.
- 8:56Traces make model improvement actionable
Malde describes a trajectory, rather than a token, as the economically valuable unit of work. He says companies need to log agent traces, make them replayable for evaluation, and modify training and reward algorithms; he cites LangChain, Braintrust and Raindrop as tracing tools.
- 13:05Group-size-one learning cuts setup
Malde says standard GRPO compares multiple rollouts of the same task, often eight or 16. Trajectory is developing a group-size-one approach that learns from a single rollout across distinct tasks, which he says can remove the need for a simulated environment. For rewards, a meta-grader can use labels on 100 representative traces to propagate criteria across 10,000.
- 16:19Companies can train and deploy models
Malde says Trajectory’s platform lets companies bring in agent traces and their criteria for good performance, then train, evaluate and deploy in one place. He names Harvey, Rogo and NVIDIA as users. He says Rogo trained a finance model that was several times more efficient in token use, cheaper and faster, and sketches a future where trained models handle 90% of tasks while a frontier model handles 10%.
- 20:18Curiosity must stay aligned with people
Malde says agents do not yet accumulate experience: LLMs predict the next token but do not model the transitions between states. He sees curiosity as identifying unknown transitions and learning by testing them. He distinguishes human-guided continual learning from recursive self-improvement, which he describes as putting a model in a black box to improve itself.
- 24:16Research and product must meet
Malde says Trajectory brings together people with research, product and infrastructure experience, and has productionized its training code with microservices and Kubernetes. He says product builders will eventually steer self-improving models and harnesses like gardeners setting trellises for products that grow on their own.
- 27:27Continual learning could widen AI adoption
Malde says the broader ecosystem needs infrastructure such as sandboxes, and compares agent learning to recommender systems that improve from user preferences. He argues that continual learning could help companies specialize models and harnesses for more workflows; he says only 1% of enterprise has adopted AI and wants to move that figure to 99%.
Published Nov 5, 2026.
Transcript · 27 min read
Training Data, hosted by Pat Grady and Lauren Reeder, with Ronak Malde.
Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.
A lot of our philosophy in product actually comes from Apple. In the early 2000s, you had a computer that had access to the internet, and you had a phone that you could make phone calls. And I think no one really thought to put the two together until Apple said, okay, what if you actually put the internet on a handheld device and combine the two? And I think by doing this, basically putting training, eval, and deployment in one place, making that just dead simple for every single person to do. And then you can close the loop and basically now upload that model back to production, gather new traces. It suddenly is this magical paradigm that I think every single company is going to adopt very soon.
thank you for joining us it's a pleasure to have you we would love to dig into a bunch of what you're building how you see the world of research and where you think the future is going um maybe to start, you guys spend a lot of time thinking about continual learning. Can you define it for us? I think you've also said this field is in a state of pseudo-continual learning. What is real and how will we know when we get to the real thing? That's a great question. So when I think about continual learning, I actually think the human brain is a really good model for where we should aspire to get to. In some sense, we are true continual learners. We learn online, we are constantly updating iterating and i think the world we're in right now with agents and ai is kind of like you said pseudo-continual learning when i was growing up i uh my family used to do a bunch of road trips right and uh i don't know how you guys did things but we would go to triple a uh back in the day and like they would give us a bunch of these maps they would like highlight everything uh and then we'd go on our road trip uh and like if there's any point where we wanted to change our route, we'd have to either call AAA or go to a physical location. And I feel like
that's the world we're in right now, where we have these amazing agents, they're very intelligent, we have these mapping algorithms, but you can only consult offline a couple of places. And anytime you want to update your course, you have to just do an entire mass batch update. What the first goal and what we're building at Trajectory is first the ability to just democratize that access, right? To even just training these models and customizing them for what people want i think of that as kind of like map quest uh where suddenly it's still offline but we at least got the ability for everyone to not need to go to AAA, but get the ability to map these algorithms.
Read the whole transcript
But the final stage of continual learning is something more like Google Maps, where every single time you're driving, every route that you make, it is updating. It's learning aggregate what the fastest route is and where the traffic is. That's the world I want to get to. That's a really good analogy. So recently we had Rich Sutton and David Silver on the show, and both of them kind of made the comment that continual learning isn't really possible with today's architectures. I'm curious in your point of view, what's doable? What's not doable? What's the state of play? Do we need major architectural leaps to kind of bring this to fruition? Where are we today? Yeah. So it's funny because when you think about a transformer and LLMs, I feel like a lot of people have seen maybe Carpathia's tiny GPT video or something where you basically have the stack of attention blocks and transformers that are built on top of each other. And a lot of LLM scaling has just been stacking these transformers on top of each other. But when you look at like any paper, like DeepSeq v4.1 Flash that just came out recently, their architecture is actually kind of lobotomized or changed a lot of that base architecture, where you have things
like fast and slow memory within the weights. You have skip connections that are going between layer two and layer 50. And so I think these architectural changes are actually already happening within the base models. And it will be a key component, I do agree, for continual learning. Where I want to sit is basically leverage all of that technology and research that's happening at the architectural level, but then actually connect the dots to the real world. Because at the end of the day, the real value of any AI progress, but certainly continual learning, is how are people actually using models in the real world. And I think companies have this idea of basically what good looks like in their mind. And the hardest part of this entire loop is then translating that to actual improvements for the model. I think that's where that missing layer is.
Yeah. So how do you connect the dots to the real world and how much of that is product or productizable? And how much of that is just handholding and helping people
How did you pick the name and tell us how it fits into what you guys are doing?
Yeah, so it's funny because the world right now is very focused on tokens. And I think a lot of people view tokens as a unit of work. But if you think about the way that we interact with other humans, I'm not counting the amount of words that you're actually talking about, right? And with co-workers, we, a lot of the time, are thinking about a task or a trajectory. And that's the unit of work that I really want to get to. In RL, it basically is synonymous with a rollout or what you would actually do in a training stack. And with continual learning, you have an RL training stack that's like a rollout engine and a trainer.
A rollout engine is just like in a simulation, basically like doing a bunch of rollouts and doing these trajectories. In the real world for continual learning, your rollout engine becomes just the inference in the tokens in the real world. And now suddenly your trajectories are the tasks of economic value that are happening. And so I want this basically paradigm to propagate where the unit of work that we... think about is a trajectory. Beyond that, I think for also people that don't understand RL, it is also super nice that we are basically letting companies just own that path and really grow on their own. So I think it has this nice double entendre.
I like this idea of a trajectory as like a unit of economically valuable work. You've also a couple of times talked about what does good look like, which kind of sounds like an eval or a benchmark. What are the core primitives for continual learning or what are the key ingredients that kind of go into this?
“you need to be able to trace all of your trajectories that are happening in production.”
To your point on abstracting away some of that, what has been abstracted thus far? What's next? What's kind of further out? How does this product come together over time?
“we can actually learn from just that one behavior or that one rollout.”
That's the world I want to get to. And that's where we're pushing that research direction. Very cool. So you have a product that you just released to do some of this. Tell us about it. What are you releasing the product and who is it for? yeah up until now it's taken a massive research team a ton of time and a ton of compute in order to train these models what we wanted to do is just make the easiest and most intuitive interface to train models and customize them for what you care about and so trajectory the platform that we released is basically a very simple system where you can take your agent traces and what good looks like it's a one click train interface that just takes a couple agent prompts and then can eval the model and then deploy it all in one place. And I think the really magical part about this, a lot of our philosophy in product actually comes from Apple. And when in like the early 2000s, you had a computer that had access to the internet and you had a phone that you could make phone
calls. And I think no one really thought to put the two together until Apple said, okay, what if you actually put the internet on a handheld device and combine the two? And I think by doing this of basically putting training eval and deployment in one place making that just dead simple for every single person to do and then you can close the loop and basically now upload that model back. to production, gather new traces, it suddenly is this magical paradigm that I think every single company is going to adopt very soon. Very cool. Who is the product targeted at? Like, who is the ideal customer for you guys today?
Yeah. So what we want to do is make this very simple for companies that have already their agent harness in order. They've spent a lot of time there. But that next step of basically modifying the most valuable part, which is the model, making that super, super easy. So right now, what the product is meant for is mainly AI native companies like Harvey and Rogo and some larger companies as well, like NVIDIA are already using the platform. But the next step after this is I want to also move beyond just the model. So another important piece is the harness as well, because I think the model gets you efficiency.
It gets you a model that has seen a task a hundred times or a thousand times over and making that faster, cheaper and more reliable. What a harness gets you also is new capabilities. It's just something that was not possible before. And I think a lot of larger companies that are adopting AI, they also haven't done their work to optimize the harness for subagents or for basically squeezing out that last mile, even out of the frontier tokens. And so the eventual world is maybe you use a frontier model like Fable for 10% of tasks. And then for 90% of the time, you can smartly route to a model that you've trained that is far more efficient and also gets you better margins.
Tell us some stories about some of the customers you've worked with so far. You gave us a little hint at Harvey, but maybe a couple more examples. Yeah. So one really interesting one here is Rogo. And I think Rogo is taking a very interesting paradigm to agents that I think will become the future. So right now, the way people use agents are very much like synchronous, right? You prompt the agent, it goes and does its thing, and then it comes back. What Rogo is very interested in is not just like this kind of verbose model that will go and basically like spit out a bunch of garbage, right? What they want is just like a very concise and clear finance agent that basically like does the work that you care about.
It can autonomously do things on the side, but then it comes back and is very efficient. And so what the goal of Trajectory is, is to basically say, okay, Rogo, what good looks like for them is they also want the idea of intelligence density. They want the idea that a model can do the same task, but do that far more efficiently than the frontier models. data to do that. And so with Trajectory, what they were able to do is train a model that is several times more efficient in terms of its tokens and also happens to be cheaper and faster, but then is also able to gain and climb on the things that they care about.
“you need to allow these models and build the algorithm so that it keeps humans in the loop.”
You guys have done a bunch of unique things across product and research at trajectory to bring these two halves of the whole together. Tell us more about what makes building at trajectory unique and who are some of the people you bring together to make it happen? Yeah. So the cool part that we are sitting in is the intersection of frontier research and product, where we have to modify the core RL algorithms in order to get to a place where you can actually learn from the world, but also package that into a really beautiful and clean interface for people to train these models. And so some of the research team that we've built had contributed to, for example, the co-founders to Gemini 3 and Gemini Robotics, other folks to GPT-4-0 at OpenAI, and building Muse infrastructure at MSL.
But then on the product side as well, having people that have built scalable infrastructure at Stripe for payment SDKs. Also, we have the youngest design engineer at Netflix. These sort of people, and I think the intersection between them, is where the magic starts to happen. So the way we've actually structured the team, normally you have a research stack that's the spaghetti code base in one corner of the company. And then you have the product that's kind of well built out. What we've done is actually productionized the code base for training as well. So this means microservices, Kubernetes scaling, auto scaling of every single part of the trainer, the rollout engine.
And then on the design side as well, our design engineers are kicking off training runs all of the time and understanding deeply what the research experience is. I think this is a really interesting and cool place to be in. A lot of this is also only possible because of Frontier Coding Agents. Actually, most of the time now, we are spending our time whiteboarding, and we'll spend a lot of time just arguing about system design. And then the actual implementation is just a software factory of agents that will go in and actually implement it. So building a product here is not easy. What is the role of a product builder in a world of continual learning? I think as a product builder, you should be thinking about how you actually take these traces and experience in the real world, and start to shape and improve the product autonomously. So we're starting with the model, right, for trajectory. And I think there's some interesting work to do there.
Eventually, that will be the model and the harness and basically self-improving both of those. And I think once you have the harness and products are kind of becoming more AI native, and so the harness is kind of becoming most of the product for a lot of these companies, right? Well, and if you have the right to kind of self-improve the harness and then the product, then you have the right to kind of improve and kind of self-improve the entire company and I think there then products and product builders start to become almost like a gardener like you have your plant and it's going to grow autonomously on its own and your goal as a product builder is to basically just set these trellises in the ground of here's where I want to shape my product and I think this is gonna be a very exciting world where you can just put a product out to a couple hundred users it's gonna grow on its own and now we're gonna have basically you You just imbue your taste into where the company grows.
So I think that's going to be a very exciting world and companies should kind of start building towards that. For somebody listening who's thinking, boy, trajectory sounds like a cool place to work. What's the pitch? Why might somebody want to work a trajectory? So I think there's a couple of things. The first thing is continual learning as a paradigm is the next learning paradigm. Any massive company that you can think of of the previous age, so let's take Google or Amazon or Facebook, they were all doing continual learning, basically. If you think about recommender systems or ads, the idea is taking all of these preferences from the real world and then creating online systems that can learn and get smarter the more they're used. And I think for agents that hasn't existed yet and that's exactly the place that trajectory I want to be in. We're starting with just v1 of the product which is a very easy way to train models but in aggregate I want it to be a massive recommender system of every single agent should be attuned to what companies want and what every single user eventually wants and I think that's going to be a fundamentally exciting world to get to. There's a lot of work that needs to happen.
And I also think Trajectory is not the only company in the ecosystem that needs to work really well to make this happen as well. You need like two blocks down from us is modal at our office that works on sandboxes. And that powers a lot of the infrastructure that we have. And two blocks down from that is Thinking Machines also in SF. And I think there's like all these sort of companies that need to be built to basically power all the infrastructure to make continual learning work. But I think the product interface is where a lot of the exciting work can happen. Yeah. One of the things you said when we were first getting to know you many, many months
“we get to a world of specialization where everyone can basically customize every single part of the harness, the model, and that's where we get to a world of abundance.”
Yeah, thanks so much for having me. Thank you.
