> Markdown version of [/videos/1526-genai-unpacked-beyond-basic?t=1014](https://www.wearedevelopers.com/videos/1526-genai-unpacked-beyond-basic?t=1014). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# GenAI Unpacked: Beyond Basic

Are your language models just chatting, or actively executing code? Learn to map raw human intent directly into system-level operations using Semantic Kernel and advanced function calling.

- **Speakers:** [Damir](https://www.wearedevelopers.com/@damir)
- **Event:** World Congress 2025
- **Published:** August 20, 2025
- **Duration:** 27:12
- **URL:** https://www.wearedevelopers.com/videos/1526-genai-unpacked-beyond-basic

## Summary

Moving beyond conversational chatbots requires treating language models as active system components capable of routing natural language—regardless of the user's native tongue—directly into machine execution. Utilizing frameworks like Microsoft's Semantic Kernel allows developers to map raw human intent to system-level operations, such as triggering or terminating PowerShell processes via conversational input. This approach shifts the paradigm from treating AI as a pure knowledge engine to deploying it as a functional syntax translator that understands operational context while adhering to built-in content filter rules.

Beneath this abstraction layer, system optimization relies heavily on understanding byte pair tokenization and high-dimensional embedding models. By translating text into numeric architectures, platforms can calculate the cosine similarity between inputs to cluster user intents or power sophisticated recommendation engines. However, a critical architectural realization is that semantic similarity is purely mathematical and entirely detached from factual meaning; logically contradictory statements map closely simply because they share linguistic patterns. Furthermore, as developers implement large-scale storage in data stores utilizing native vector search—such as upcoming SQL Server releases—they face significant compute challenges. Query performance in raw vector databases degrades at an almost linear rate under heavy volume, requiring specialized optimization to prevent severe application timeouts.

The deployment of robust AI-driven agents also demands demystifying the perceived intelligence of language models. By analyzing token log probabilities and tuning the temperature parameter to zero, it becomes evident that LLMs act as deterministic probability engines, statistically chaining semantically coherent text sequences rather than reasoning. This mechanism natively empowers advanced function calling architectures. When executing frameworks within tools like Azure AI Foundry, the model calculates prompt embeddings against registered application plugins to extract parameters and trigger compiled C# code. Consequently, in modern AI workflows, meticulous source code documentation transforms from a developer courtesy into a functional runtime necessity, as autonomous models rely explicitly on human-authored metadata to map unstructured queries to executable logic.

**Keywords:** generative ai agent development, microsoft semantic kernel integration, natural language command execution, byte pair tokenization encoding, tiktoken cost token extraction, cosine similarity semantic matching, multidimensional vector embeddings, vector database query bottlenecks, sql server native vector search, llm temperature parameter tuning, text generation log probabilities, large language model function calling, azure ai foundry deployment, c# programmatic parameter extraction, source code documentation for ai

## Chapters

1. **Controlling machine processes using natural language AI agents** (00:05) — How large language models map human instructions directly to executable shell commands.
1. **Understanding text tokenization and cost calculation via byte pair** (06:26) — How the byte pair algorithm encodes text into numerical arrays to improve indexing performance and track API costs.
1. **Calculating semantic similarity with embedding models and vectors** (10:36) — How embedding algorithms convert textual data into dimensional vectors to evaluate conceptual proximity using cosine distance.
1. **Storing and querying vector data with native database features** (16:54) — How modern relational databases index structural vectors natively and the scaling challenges associated with high-dimensional search latency.
1. **Adjusting model creativity through temperature and log probabilities** (18:59) — Why tuning token probabilities influences the deterministic capabilities of large language generation.
1. **Implementing dynamic function calling with semantic kernel agents** (23:55) — How integration agents calculate intent embeddings to route prompt criteria strictly to registered application functions.

## Related Moments

- [Understanding language models and autonomous executing agents](https://www.wearedevelopers.com/videos/1725-wearedevelopers-live-build-a-multi-ai-agents-game-master-with-strands-our-weekly-web-finds) (from "WeAreDevelopers LIVE - Build a multi AI agents game master with Strands & our weekly web finds")
- [Exploring the core architecture and components of AI agents](https://www.wearedevelopers.com/videos/1510-on-a-secret-mission-developing-ai-agents) (from "On a Secret Mission: Developing AI Agents")
- [Leveraging the comprehensive generative artificial intelligence stack](https://www.wearedevelopers.com/videos/969-make-it-simple-using-generative-ai-to-accelerate-learning) (from "Make it simple, using generative AI to accelerate learning")
- [Architectural components of a generative AI agent](https://www.wearedevelopers.com/videos/782-building-real-time-ai-ml-agents-with-distributed-data-using-apache-cassandra-and-astra-db) (from "Building Real-Time AI/ML Agents with Distributed Data using Apache Cassandra and Astra DB")
- [Defining modern generative artificial intelligence and agent applications](https://www.wearedevelopers.com/videos/1512-building-ai-applications-with-langchain-and-node-js) (from "Building AI Applications with LangChain and Node.js")
- [Understanding core parameters and mechanics of large language models](https://www.wearedevelopers.com/videos/1512-building-ai-applications-with-langchain-and-node-js) (from "Building AI Applications with LangChain and Node.js")

## Related Articles

- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [Graph and AI Trends 2026: Why Is AI Running but Not Yet Delivering?](https://www.wearedevelopers.com/magazine/680-graph-and-ai-trends-2026-why-is-ai-running-but-not-yet-delivering)
- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)

## Related Jobs

- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Senior AI Agent Software Engineer (Go, Python) (m/f/x)](https://www.wearedevelopers.com/jobs/48277-senior-ai-agent-software-engineer-go-python-m-f-x) at **Dynatrace**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/588393-machine-learning-engineer) at **Twilio**
- [AI & Machine Learning Engineer (all genders)](https://www.wearedevelopers.com/jobs/48217-ai-machine-learning-engineer-all-genders) at **msg**