Nupur: I'm Nupur. I work with Kodo. Uh at Kodo we do agentic reviews. Uh I have a background in dev sec ops. So I'm coming from an industry where everything was deterministic. The pipelines they run they crash. If they crash, we fix them. Uh to a place where we are doing agents where nothing is deterministic. So in my last few years I have learned where and how agents fail, what are the learnings and today I'll be sharing some of my learnings with you.

Nupur: So um if you see the evolution of agents, it started with static prompts where it was a 4K context window and we tried to put whatever was important or whatever we deemed important and the AI models will process it and provide you with a result. Right? When we started with that that means that it was on us to tell LLMs what do they should they should look into. That means if we provide wrong inputs, we might not get proper results.

Nupur: And then we thought maybe if the context window grows, if the context size grows, we can do better. We can have more inputs. And uh we started with agentic workflows. So we created an agent. We get that get them tools like search tool to go into search into documents and do something uh as a command. then again look into the search and do something which again created kind of a loop where the tool does not know where to stop. It thinks like I need more inputs again going back and back. It's a loop to improvise on that.

Nupur: Uh nowadays multi-aents is becoming more popular. Create multi-aents do a lot of stuff together. When we see it like that we have a lot of agents working for you. a security agent trying to figure security concerns, a review agent trying to review the tool, a coding agent trying to fix things. Now again, the more the tools, the more issues you have. Not every agent understands and they have clash in their understandings where you don't get into the results.

Nupur: So what do we learn from here? What uh we see is context is not a problem. Day by day the models are coming where you can dump a lot of context, a lot of data. But does that make sure that the results you are getting is smart enough to give you everything or smart enough to decide what's important? If you see the current LLM models, we see a pattern where it takes the initial inputs you provide, it takes the last inputs, but the in between context is basically removed. So they don't focus on the in between context. Agents look at the starting point, end point and try to provide you the results. This is like a U curve where some of the things from the start, some of the things from the end make sense but whatever you are providing in between that that is not taken up. Yeah.

Audience: How do you know this?

Nupur: This is something which we are working on and we are actually benchmarking things. So when we create agents we try to see this is the context we provide to the agents. Does it take this context into effect and also give us the results. So we are working with multi-agentic architecture where for each of the task we do do for code reviews we give the task to an agent and say okay give us a result.

Nupur: Now every time we for example code reviews we try to see can we give all the context can we give the whole codebase for example and see if we can get a results but we see that that whenever we start working with that the initial prompt or the initial goal which we start with that is in focus. If we give something at the end as an input that is in focus but all between context like I have Jira I have MCPS can you look into that the LLMs try to get rid of those things and purge them to get make sense by themsel.

Nupur: So to have this or to make a way out of this how we deal with is creating strategic solution for context optimization rather than dumping everything to the models and asking them to be smart enough to find out what is more important. Uh we usually try to see okay what we can do to make it uh better context for the model. There are lots of solutions in place if you see currently and context engine is a buzz word like everybody wants to create context engine and everybody wants to uh provide that but context engine is like a bouncer right so your high-speed car is going and it acts as a bouncer and tells you this is more important now if you have a large messy code base it makes sense to create a context engine because it creates a search pattern it creates ranking logic so that whenever you ask for a task it looks for those rankings and say this is more important for you take it and work with it.

Nupur: The problem is the indexing part takes moderate effort but the scaling is a challenge. Like if you start talking about 600 repositories or 700 repositories the mapping and the indexing starts to slow down and it becomes again unpredictable to find or create a context engine if you are not actually into making context engine only.

Nupur: There are lots of areas where agent can get more context instead of investing highly on context engine hierarchal summarization where instead of creating or going through everything a summary is created for each file and folder so that when the agents try to find they can try to read the summary and see if that is more important to us or not can be a good one. The only thing is that you need a lot of LLM processing. every time a file is created or changed some of the agents need to go and create a mapping for that. So it's a high upfront um context of uh LLM processing that is needed.

Nupur: Another way is knowledge graph. Now knowledge graph is complex but it works wonder when you have logical dependencies. For example, you have one file which impacts another file which again impacts other file. You can create a graph DB hosting. It is the initial input needed by the developer is very high. It takes a lot of time to create that. But if you have complex logics or you have dependencies on multiple repos that works wonder for me.

Nupur: I think for most of the task if you're not a product company but if you're building agents for yourself or your processes iterative retrieval works really good because instead of even creating a summary it creates kind of an index. So it's like a library card which you give to your agents and say this is the topic if that is relatable to you. You can look deep into the code uh and uh look for the results. Again it has quite uh cost impacts but you do not have to invest a lot of energy. The input uh required by the developers to provide to the LLM is low and it provides better results.

Nupur: There is also option of selfcorrection where you ask uh the LLM to do something and there is a critic node which looks and say if that is relevant to your initial goal or not. In that case if the context is lost uh you can again ask the agent to do it again retry it again because the critic note said this is not the right way. It takes a little bit more time because it adds a latency of running the agents and again but it does not require a lot of input initially from the developers to create something.

Nupur: Another challenge uh which I have seen people getting into when they create uh these um agents is the orchestration paradox. Now what it does is that now LLMs are becoming more and more smart. So when you give them the task they think like okay I should use this tool uh maybe I can do better I should research more on what should I use. It goes into a loop that instead of actually looking into solve the problem, they look for the method to solve the problem. They hop on from one method to one another method and most of the API tokens are wasted on finding a way to do it rather than doing it. So you will just go on the research mode. For example, if you use Opus latest and greatest, they will try to see what is the best method to do it and challenging themsel again and again. Maybe not this, another way, another way. And it just goes into a loop of doing trying to do something rather than doing something.

Nupur: To resolve this uh we worked with 80/20 hybrid approach. I think this is one of the most interesting uh outcomes I have seen or the way to in resolve this infinite loop. What our teams are doing is giving the latest and the greatest models or giving the agents power to research 80% of the time. So you give them the goal and say okay try to do whatever you can but the 20% of the task where you need final validation you want summarization that are not something which is free flowing that is more restricted those are more hard gates for example if I get X results I want Y it's more deterministic so that the research which is coming from the 80% can be lowered down now when you see you can always say that the 80% tool can still go on and go into infinite loops.

Nupur: We have mechanisms to work on that. For some organization do they do counter mechanism where after four or five counters you have to work with whatever was the last results. For some of them they have timeout counters that after 5 minutes whatever is the last tool or whatever is the last decision you work with that and then go back if the results are not good. But you can restrict that 80%.

Nupur: But in short, if you are using anything like discovery or you're trying to see which tool to use, you're trying to plan those 80% uh research models are really good. But if you are again trying to create a summarization, you're trying to see okay this is the research I have got. Now I have to make a result out of it. The 20% works really well. Now for 80% usually you use uh high reasoning models latest and the greatest but you don't need a high reasoning model for the 20% because those 20% things are doing deterministing they you are actually telling them what is needed for example the critic node which we talked about they don't need to research they don't need to find out what is the best thing to do they just need to see what was your goal what was the result you are trying to achieve and uh how to provide or how to summarize ize that for that also things like if you think about what would be the next possible action I have this result what should I go and look for that are things can be done by the 80% dynamic models whereas I have all the results from the 80% models but what is the proper way or what is the proper uh results the the user is looking into those kind of decisions can be taken by the 20% model.

Nupur: Finally, uh this is again an interesting failure which we have seen where as the context grows, teams think okay we can do everything with one agent because the context window is quite great right we can put everything we can ask an agent to uh do uh the testing part we can ask that agent to do review part we can ask the agent all the kind of things because the context is same and they can provide us the results that make sure that the when the agent is going forward ward it get overwhelmed with the inputs and again it tries to start losing uh what was the original task. So maybe you give four tasks to the agent and somewhere down the line it focuses on two tasks. So you get great results for the two task but the other two just just get lost in the middle.

Nupur: For those particular purpose, we have something called mixture of agents. And that's that's where you hear a lot of buzz about multiple agents or multi-agent tech architecture where you create instead of one big agent, we create issue expert agents. We create small small agents which are doing great in a specific task which they have provided. Now to build on top of that each of the agent come up with their own interesting ideas or results. How to make sure those results combine and make sense together. Because for example I am trying to search for a vacation. I give an agent to find the best hotel. Another agent try to find best flights. But all three of them gives me different result. The uh hotel is in Greece. uh the flight is from Amsterdam to um maybe Portugal and everything just doesn't make sense right so for that particular purpose there's a concept called a judge agent what it does it it tries to get all the results and see if they can make sense together so now you are doing all the greatest things from different agent getting the best results from their part but a judge agent help us to combine these and make one sense out of it instead of getting so many things which doesn't make sense together.

Nupur: Something similar is implemented by us. So this is our architecture uh codos architecture where for the code reviews we are using the same uh formula. So as part of a PR review we have a context collector which actually goes and collect context from the PRs it could collect context from the context engine. it uh collect context from the tools but then it does not start working and uh giving you the reviews. It actually bifocates all the context it has provided and pass it on to different agents.

Nupur: Now what these agents do is basically specializing in what they are supposed to. For example, there will be a security agent trying to find security flaws. There might be a agent trying to code. There might uh code differences. There might be an agent trying to find the Jira issues. Once all these agents give us back, a judge agent actually looks for the results and say okay these are interesting enough but is it relevant to you? They can again go back in the context engine look into the PRs and see out of the 10 things which is provided by you how many of them actually make sense for your thing. So again refining the results uh to make sense to you. Yeah, I think that was it from my side. Uh, any questions?

Audience: Yes. Um, in practice, how do you let the swarm communicate with each other?

Nupur: Uh, you are talking about the agents.

Audience: So, you have agent A and agent B and the judge. I can imagine they write to a file system or they have some kind of tool proprietary.

Nupur: uh we ch we use uh longchain at the bottom uh and that is being used to communicate and build the infrastructure for uh different agents.

Audience: Do you know what lang chain uses for that like just collects the responses and then shs it back into the prompt of the next agent?

Nupur: Yes. Yes. Yes. Yes. that that's what so what we do is we try to uh get the results and create a prompt for the next agent and if it's multiple things again there is an agent just to collect the results and create a better prompt uh which is refined for the next agent

Audience: have you thought about a calibration step for each agent

Nupur: when you say the calibration can you tell me more

Audience: calibration right so when you do a code review with an agent right need At least what I heard today is doing some calibration that you actually tell them what is good and what is bad.

Nupur: Yes. So when you say it like that uh and let me know if that makes sense for your question or not. We do calibration in a form we check what we have as a context. So for example uh when we get the code reviews LLM does not know what is important for you or how do you work. So for example an LLM when they gets input they get input from healthcare industry they get from retail industry they get from finance industry and all of them can use same Java framework in different ways or different uh things are important for them the rest of them doesn't make sense.

Nupur: So what we do is we give you two different options to tell the agents how to perform or what to work on. One part is we give them the PR history. So we index all your PR and see when was the last time something like this was identified and compare the current version if that is

Audience: you do that in the context standpoint.

Nupur: Yes. Yes. Yes. Yes. We do that. So the changes we you make to the code we look into if we can see something similar in the past that again is uh transfer to the context twice. First is when we are actually giving context to the sub agents to find things for you and another time to the judge agent. So that when I get 15 different recommendations for a code review, my judge agent can look into what was there before, how did your reviewer commented, how your developers commented and based upon that decide if that is worth providing to your developers or not. And the other part is

Audience: and this happens for every agent.

Nupur: For every agent yes,

Audience: right, you don't share the context between the agent, right? So you have a specific context by the way.

Nupur: Yes, we are trying to resolve that U part that instead of dumping everything to the LLM, we uh take the part which is more important, we use a context engine for that. We take the part which is more important and only provide that particular part to that particular agent.

Audience: But but then for me it's not clear how you bridge the gap rather than say you have quality agent and you have agent framework specific coding radio agent right and then you basically you only as I understood you only share the specific information to each agent right and then basically each agent runs atomic autonomously Yeah.

Nupur: Right.

Audience: And doesn't have a full picture, right? At least when let's say as a human, right? And if you do code reviews, it's always good if you have a full prospect, right? I would say like that kind of methodology would it works for simple things like does it use linting? Does it use I don't know are tests implemented? But when you think about at least I think when you think about architecture for example to make architecture decisions um that covers security because everything is a balance and and uh you have to wait that somehow so how do you solve

Nupur: yes so I think if you look into the older version of code reviews you should have you used to have a senior engineer who knows your code who knows what kind of packages you're using and they can comment on if the developer has done something uh similar to what you're used to or did something uh totally weird, right? Then you used to have a security person who used to see if you are providing all kinds of security, you are not hard-coding your APIs or you are not putting any SQL injection. So all those kind of things the security experts know. On the other hand, if you are working with ISO compliance or sock to compliance, there might be an auditor who might ask the uh team lead or the senior engineer is your code uh being um you know logged is it logging the changes and so on.

Nupur: So previously as well there were lots of people having specialized knowledge looking into those kind of areas specifically. Now when the context is provided it's always like these are my security concerns which I always have to look into. These are my architectural concerns. An architect might look from the architecture perspective. We can do that something similar uh with the agents as well because for example architecture security concerns we have a web portal where architects can provide their guidelines pro uh compliance people can provide their guidelines and an agent can look into all those guidelines and say is it validated or not.

Nupur: So if you see the initial uh picture the context collector knows everything and then it provides relevant context to the agents. So you you enforce basically your customer to upload these kind of documents, right?

Audience: So you you enforce basically your customer to upload these kind of documents, right? Is that a kind of a requirement then or because I mean the system will have completely let's say different result right if you don't share them

Nupur: exactly so that's something which depends upon organization there are some people who say we don't work with any rules or regulations so just give me out of the box that also means that don't expect the agents to find something very specific to your working until and unless we have certain PR history because then again the PR history kicks in and tells you what is relevant even though when you don't provide

Audience: I'm not sure if the the PR history is really the best source but I think it it can be one it can be one of it can so

Nupur: that's why there are various uh sources right so it's PR history it's your resource it's your and somebody's cooking food at my home yeah but it it can be uh one of the source and that's why

Audience: and my question is like yeah I mean of course It can it can be right but yeah at the end you need to decide let's say also in the jury right how much you weigh right so like the new documents for let's say the engineering principles architecture principles and compare compare them to let's say your your mer yeah yeah right they could be completely out of balance right

Nupur: um it depends it depends so again uh I think It's if you look into from one perspective it's difficult to decide but if you're getting the context from many angles for example PR history that's one part but when you do the compliance and you tell in the compliance portal this is really important. So we have various uh segments of it's an error, it's an uh a recommendation. All those kind of things adds weight to a feedback to say if it's good or not. And every time your developer expect accepts a suggestion, it gets more weighted for the next one. If it does not uh uh accept the suggestion, it gets a less weight. So it's all about indexing and making sure those weights are managed somewhere.

Nupur: Uh two ways. One is uh by when you when you give your recommendations does your developer actually accept it or not we uh index that. Another way is uh from the past PRs we try to find out similar issues identified and if your developer actually implemented them. So for example some people are used to hard-coding their uh API keys and I literally had a tough um argument with the developer but this is how we do it. No this is should this should not be the way. Yeah but nice thing nicely match also what I meant right so if you look in the in the history that happened doesn't mean it's it's good it's yeah and and that's the way where the system tries to tell you this is not good this is not good and then it's up to you

Audience: only if you provide the guidance right

Nupur: no it so there is something called bug fixes and there is something called rules so if you provide them as a rule it will get uh highlighted doesn't matter if you want it or not. And then there are bugs where agent tried to tell you there's something wrong and if the reviewer also agrees with it and did not implement a 10 times the reviewer might get it less weighted and give you

Audience: cool. Thank you so much.

Nupur: Thank you.