The Self-Improving OSS Agent Stack — Marc Klingen, Langfuse

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: The Self-Improving OSS Agent Stack — Marc Klingen, Langfuse · Channel: AI Engineer · 16:36 · Watch on YouTube

Topics: AI agents, AI evaluation and benchmarks, Data, databases and embeddings

Summary

Marc Klingen describes an emerging open-source stack for building self-improving AI agents, combining production tracing and monitoring with offline datasets and evaluations. Teams can use AI to analyze production signals, propose changes to agent implementations, and test those changes against maintained datasets, while people retain control over goals and evaluation criteria. A demo shows a change-log writing agent identifying problems with clarity and internal jargon, then proposing and testing improvements. As agents process more traces, the approach also increases the need for scalable, cost-effective data systems that retain production data over time.

Key points

Effective agent development connects online production traces with offline datasets and evaluations, creating a feedback loop between real usage and testing.
AI can reduce manual work by identifying failure patterns and proposing changes to agent implementations.
Teams are beginning to use AI to suggest updates to datasets and evaluation criteria, but human review helps define the boundaries for optimization.
People should set goals and decide which changes matter, while lower-level, repetitive tasks can be automated.
Production feedback can include explicit user complaints as well as signals such as human edits or approvals.
In the demo, a change-log writer was factually accurate but lacked clarity and used internal jargon; the agent proposed tests and implementation changes to address this.
Candidate agent changes can be back-tested against an updated dataset and compared with the existing implementation before deployment.
Agents that analyze large volumes of traces make data access more read-intensive, increasing the importance of scalable storage and long-term data retention.

Chapters

Start	Chapter	Summary
00:01	Why agent development is changing	Klingen introduces the shift toward self-improving agents and explains how increasingly capable models make higher-level automation possible.
02:06	Connecting production data and offline evaluation	He explains why agent teams need to link production traces with datasets and evaluations, and how AI can help automate this feedback loop.
04:13	Where humans stay in the loop	The talk distinguishes between AI-proposed improvements and the human role in reviewing datasets, evaluation criteria, and optimization goals.
08:23	Balancing automation and agent quality	Klingen compares manual and fully automated workflows, arguing for automation that reduces tedious work while keeping people involved in setting direction.
10:26	Demo: improving a change-log writer	A demo shows an agent using review feedback to identify unclear writing and internal jargon, propose evaluation updates, and test a revised implementation.
15:05	Operational lessons and data needs	Klingen summarizes how teams can run improvement loops on production signals and why this calls for scalable systems that retain data over time.

Quotes

you need to bring the online and the offline together — 02:06

you want to be in that loop involved because you set there by the direction — 07:51

AI can just look so through like so much more error cases. — 09:24

we improved on like userf facing language. — 14:02

you want to really like own the data layer — 16:06

Concepts and entities

self-improving agents, tracing and monitoring, offline evaluation, Langfuse, Marc Klingen, Loopcraft, change log writer, GitHub

Full transcript (original language)

00:01 [music] Everyone, super excited to be here. Ah, okay. My laptop was alerting me that this starts now. Uh, hi everyone. I’m Mark, one of the founders of Langfuse. super excited to chat about uh what we have seen how how people like self-improve agents now and what kind of like open source reference stack uh we see emerging. I mean obviously my view

00:31 is based on working with our community of people that use language to trace and evaluate their their agents but I think I try to keep the talk mostly to to like more generic takeaways that work with whatever stacks you’re using. Um but yeah happy to chat about the different pros and cons after the talk. high level. I think this is the year uh where uh teams talk about like upleveling the the level of uh where where they operate themselves where increasingly good good models just abstract everything downstream. I mean coming from I don’t write prompts I have loops u Peter

01:02 writing about it auto research becoming really popular. So I think this is kind of where things are going on the like like producing um application side but also on the how to actually um build good good agent applications and um over the years really models have have expanded a lot in capabilities um back when when we started working on length fuse this was right when ch was released so uh where like even the simplest applications didn’t really work so for example we built like github issue to pull request automation similar to what devon or kurs are building but back then

01:34 we could only manipulate like single page applications like a single HTML file because multifile edits were too complicated. Since then all of these things really accelerated a lot and this really made made loops not possible. At the same time a more like reference tech emerged of how great AI agents are built because building agents is different from building applications. It’s like an like a link of like tracing and monitoring like online how users really use an agent application and then like an offline component of how to build data sets how to experiment uh regarding

02:06 like making changes to these agents and then evaluating offline whether things work then deploying new to production then learning again for real users use the application uh so it’s like this this kind of loop where traditionally I mean you have like obsibility and the data docks here in the tracing monitoring side and you have like the ML ops toolings like MF flow weights and biases more on the like offline side where this new category emerged and that’s also why we built length fuse because you need to bring the online and the offline together as as otherwise like you either benchmark on data that’s inaccurate so the data sets they’re not in loop and like uh not in sync with

02:37 what’s happening in production or you monitoring like production data but you don’t really benchmark offline so you need to bring the the both together but this has been like a lot of manual labor because you need to look at traces update these data sets think about like new evaluators uh then make these changes create new hypothesis like it’s it’s like a lot of work going into this process and since starting working on language we try to educate teams how to how to do this process well but it’s it’s usually like the ask them like oh models got better now how can we take ourselves out of the loop to uh to automate it because it’s really tedious

03:08 uh to get things going and that’s what I want to talk about how we see how people go from manually driving this loop to using AI to to drive this loop um we use the visualization um like of for example like the loopcraft article or like many many articles in the space that like layer different loops on top of each other. How I mean at the at the lowest level you just have like this token loop of what’s the next what’s the next token that’s produced and on the very high level it’s uh like like the most abstract form of thinking where you as a human are still involved. So what we see how um we go through them one by

03:41 one high level I just want to highlight these are again possible because models improve. So uh when when we started working on length in 2023 like I mean we had github copilot so like the lowest level or the uh level two u then level three was possible like more in 2024 and now2526 is more like the higher level loops of actually agents reasoning about what what to even fix uh how to propose new fixes and how to how to verify whether they actually work. Where humans are currently still involved is one you need to align data sets of example

04:13 questions of how the agents are used. This is so like where where humans need to review um like like what what gets amended to the data set and how to then evaluate whether the agents actually work. These are the the two blue things in in loop number three. Then on the proposal of fixes because if um like your agent proposes how to change an agent, you still want to see what was changed because you have the risk of like overfitting it to um to like the data set and the failure definition. Um, so I mean usually agents find all sorts of different uh failure patterns, but

04:45 maybe some of them don’t really matter that much and you’re like, huh, this kind of use case for the agent, it doesn’t really matter that much. Like this is like out of what the agent should be actually be doing. So we don’t need to over fit our application to this random quirk that we found in production, but we want to like be in the loop to to monitor this. And now coming from this loop pattern, we try to dissect it again into more like a workflow where where we currently see AI used the most is in proposing fixes. So if we have production traces and we have like evil criteria and data sets of how to reproduce issues that we find in

05:16 production, then this how do we hill climb against the data sets? That’s like a really cool uh way to kind of like loop your way to a success with with agents because usually they’re like 10 different things of what you could be doing. I don’t know, use a different model like aggregate context in a new different way. Try whatever you find on X uh in a given week to to see whether it can like uh make a dent into improving the agent. But really like you maintain the boundary of what you optimize against. And you use AI mostly for like proposing improvements to the agent implementation.

05:46 Like over the last I’d say month, we see more and more teams also using uh like like AI and and their loops to maintain evil criteria and maintain data sets. So uh usually what goes in here is you want to like align how a data set uh looks like for an agent with what actual users are doing. So for example if we have an customer support application like support agent then the data set should be like common support questions but like users do all sorts of things with a support application. So you need to continuously keep it in line with uh

06:17 with what users are doing and you can use agents for that. And for EVAs as well if you see usual error patterns then you can reproduce the error patterns and propose new evil criteria on these offline data sets. So for example I don’t know I want I don’t want to name competitors. I want to be concise. I want to answer in the same language as the user actually requested. Um like the question in my customer support application. So you can also propose new evaluators. This is I think rather newish over the last months. um and where we see like the AI more autolooping. However, usually teams that

06:48 are still involved in reviewing these changes because they create the new boundary for how then other agents try to auto improve against it. If you give this completely out of hand, you risk like that you just create more slop where we’ve used this in like a blog post of ours where I’d say the target of an AI application is always changing because you don’t really know like usually you start with a very high level task of I stick with customer support. You stick with someone says hey in our company customers need to wait a long time for support answers and we have a lot of people employed to respond to

07:19 them. AI should be able to do this automatically or at least eight in customer support, but it’s a very wake target because people don’t even know what happens in customer support every day. So, it’s kind of like, oh, we want to automate all of it. However, then on the way of doing this, you figure things out based on like error cases of what you even need to do. And that’s like this kind of like map that you like you assemble the plane while you’re flying it. Um and how we then look at the uh at this kind of like looping graphic is you really want to be like involved in setting the setting like the the goals and how you make changes to data sets

07:51 evaluators you want to be in that loop involved because you set there by the direction and then AI can automate against it and then off of the newly implemented changes you’ll see new error classes and you can then like change cause again on the most highest level but thereby you pull yourself out of a lot of like the more manual and tedious work on the lower levels. Um, this is then the high level view of how how I would conceptualize like the different scenarios that you can find yourself in. I’d say if you implement this uh like AI workflow well of tracing

08:23 online uh how your agents are working implementing evals on how they are working um having data sets uh of example questions and then benchmarking against them you are in this era in this class like very high time invest because you need a team who really works with this data every week but also quality of the of the agents is pretty high I mean that’s what teams have been for example using length use for for the last years and this is how you can get to I’d say success with your agent application. Um, if you fully automate it, I would say you don’t really need to invest any time because you’re just like, I don’t know,

08:53 codeex goal mode your way to success or like you add higher level loops on top of it, but you also risk that like it produces a lot of tokens, but don’t really make sense. So, I think you want to be in this kind of like promised land of you don’t really need to invest a lot of time, but you’re like in the loop at the most important steps of like setting the direction of where you want to go, but you take yourself out of everything that’s tedious. Thereby you’re way less time invested than running it manually. But also the quality can be even higher because usually if you need to do everything manually you’re bottlenecked

09:24 by your own time your own like perseverance endurance uh or like patience to look through so much data and AI can just look so through like so much more error cases. This agent will be higher performing and you need less time invest. That’s the I think that’s what we were going for and how this then looks like is uh you involved at the top but also I mean you want ideally your application to feed in interesting signal. So for example when we talk about the like customer support application use case usually I mean customers can swear at the support agent

09:55 customers can be like oh no you don’t understand uh like this was not correct or like you can propose messages to an internal agent who then can accept them or change them when sending them out. So there’s like a lot of like implicit signal that can feed into your process without you as the agent developer needing to do this manually. Uh thereby you have like an like an instream of of signal that your loop can act on. Um so it’s either you your team or some signal coming in um and the lower level loops can be automated. That that’s high level how we how we think about it. And like

10:26 uh like we put together like a super quick demo application um that can that can show this. Um so for example like we as the language team we have a big problem now because our like engineering team uses AI a whole lot thus they ship a whole lot and somehow we need to update our customers uh and user base uh what is what is new every week because uh like in recent weeks like a lot was released uh and usually need to update documentation update a change log post about it on socials and historically like engineers did this themselves

10:56 because they maybe released like a bigger feature like once every month so it’s okay to spend a couple of hours. But now if they take themselves out and really ship a lot of like product, they want to ideally uh like automate also that kind of like release process because otherwise we are bottlenecked by how fast we can communicate about changes. So um we thought about like putting together like a change lock writer that just uploads like like updates a lot of like the public assets based on what we’ve shipped. Um however then the question is is like how we communicate good because if we communicate badly about what we’ve released then I mean we we kind of like

11:29 sabotage our releases if we like misrepresent for example what has even been released. So um setup is we have a merge PR, we have a change writer that uh suggests a change to documentation, but then there’s like a just like a code review step as like a GitHub PR where someone can review and either like approve on GitHub and get it merged or like submit like change requests of like what needs changes in this um in this in this change in documentation and and the change log. Then the change writer would

11:59 act on this again to update the draft until it’s finally released. And now both of these kinds of like either like approvals or uh requests for changes can get feed into like the AI obserability and eval to to serve as a basis for auto improvement of that change writer agent. So what we’ve now done is um just use a coding agent pointed at the like length use agent skills but generally I mean this would also work with other systems but yeah length is like well positioned for this to to ask like okay what has

12:30 happened in this application how can we improve it I’ll I’ll pause a couple times because otherwise it it’ll go it’ll go by very fast. So what we for example here identified that the writer uh so the chain lock rider is factually uh like very correct in how it makes changes to documentation but clarity is low. the edit ratio of us humans making ch like submitting change requests is still high and that we leak internal jargon to our customers because maybe our code has like internal comments that

13:01 we leave there for ourselves for the future but uh like our customers don’t think about our application as like I don’t know like an ingestion pipeline or as like I don’t know some kind of like batched evil Q like they don’t know this is like an implementation detail we should talk about it as like an user domain language and that’s like uh something we have identified here we’re Now the agent suggests changes to the data sets and evaluators to test for leaking jargon and um also like increasing clarity in customer domain

13:32 language basically. So um agent now closes this gap suggest changes to the data sets and evaluators and now um has an idea of how to change the agent implementation to basically fix this problem because in the end it’s kind of like just like a a problem of how we provide context how we like provide the skill that then educates this um content writer and now we can back test it basically on the updated data set to see whether the new implementation would do better than the old implementation and

14:02 I’ll stop here again where we have basically the baseline. So um basically we reproduce the issue of userf facing language. So let’s talk about features in a way of how users would talk about it. So we reproduce this kind of problem with the v1 baseline. This basically existing implementation of the agent and agent came up with a new implementation where we see the v2 candidate. We are still doing on format compliance and accuracy in the same way as we did before but we improved on like userf facing language. So we see a positive delta which sounds like a good change

14:34 now. Um so uh agent reasons about it as this. I mean this is a bit I mean internal application it’s a bit yolo. So we are like okay we are good of of just merging this and then acting again on the next change because this agent never publishes anything to actual production. It just raises PS on a repo. So it’s okay if it kind of like auto improves and then we just provide feedback again to whatever the next increment is. So um here we use our like prompt management logic so that the agent can just basically feature release this new um

15:05 implementation and uh now it just summarizes basically um how how we acted on this. So high level um this is what we have seen like most of our best users do already uh when they when they build agents like collect production signals uh track them alongside like detailed execution traces and then run agents like on a loop uh usually like a crown job like every day every week whenever you have like a batch of basically interesting user data again to then act on it. Um either it’s just user feedback

15:35 how how it’s happens in this case or internal labeling from like a production system but it could also be you ask people to annotate data like in app for example in langu we also allow like for annotations to feed into that uh into that queue and yeah that’s very high level what we’ve seen happy to give you like a more deep dive um demo at our at our booth or after this talk but just wanted to to wrap it up very concisely with how we see this this space evolves and um I think what’s what’s interesting here is it drives the need for like a very scalable data system because you’ll want to have agents really loop on this

16:06 data and like produce lots of queries. This we see how langu was historically very right intensive to like um ingest evils and traces. Now it’s way more read intensive because agents can like like chew through so much more data. interesting number one and two you want to really like own the data layer because like now the traces of like for example a year ago are interesting context and you don’t want to want them to be locked up in like like a more commercial system or something where uh you need to sample data or retain them for a long short time period but we want to like retain them for a long time

16:36 period and not sample them and yeah this really drives the need for like a scalable cheap solution and we’ve worked with many customers on this so happy to talk one about uh oneonone about this if you’re interested thank you so much for your