FOXKRIT AI TRAINING

AgentX Training Program

AI is no longer a future skill — it is today's hiring requirement. Every major company is building AI-powered workflows, and the demand for developers who can actually build Agentic AI systems far outpaces supply.

Yet, most engineering graduates leave college having never written a single line of agent code. **AgentX fills this gap.** It is a 2-month intensive program that takes a student from zero to a deployed, working Agentic AI application.

The Opportunity

85%

of new tech job postings mention AI/ML skills

3x

more salary for AI-skilled graduates vs peers

0

college curricula cover Agentic AI end-to-end

What is AgentX?

Structured training, real-world coding, and live team collaboration designed to build industry-ready AI developers in 8 weeks.

Program Name

AgentX Training Program

Duration

2 Months | 8 Sessions per Month | Weekend-based

Session Format

45 min Theory → 15 min Live Demo → 75 min Guided Lab → 15 min Debrief

Target Students

3rd / 4th Year B.Tech / BCA / B.Sc. Computer Science or related

Month 1

Training Phase — 8 structured sessions covering the full Agentic AI stack

Month 2

Project Phase — Teams of 3-5 build and deploy a real AI system

Support

Guided lectures + weekly doubt sessions throughout both months

Primary Stack

Python, LangChain, LangGraph, CrewAI, FastAPI, Streamlit, ChromaDB, Docker

Outcome

1 deployed AI application + GitHub portfolio + interview readiness

Month 1

The Training Journey

Each session builds directly on the previous one. The thread running through all 8 sessions is a single project — a Personal Research Assistant Agent — that grows in capability each week. By Session 8, it is a fully deployed multi-agent RAG system.

Milestone
S1

AI Foundations + Professional Python Setup

What you learn:

How to call an LLM API, structure a Python project, and manage API keys safely.

Research Assistant Dev:

A working chatbot that answers questions using Gemini/Groq. Just one script — but it runs cleanly, with proper folder structure and .env secrets.

S2

Prompt Engineering + LangChain + Structured Outputs

What you learn:

How to control LLM responses using LangChain, Pydantic, and prompt design.

Research Assistant Dev:

The assistant now returns structured JSON — summary, confidence level, key topics. Output is validated. It no longer breaks downstream code.

S3

Embeddings, Semantic Search & RAG Concepts

What you learn:

Why semantic search exists, how embeddings work, and how to retrieve meaning — not keywords.

Research Assistant Dev:

The assistant searches uploaded documents by meaning. Student runs the same query via keyword and semantic search and sees the difference live.

Milestone
S4

Vector Databases + Full RAG Pipeline + Persistent Memory

What you learn:

How to build a complete document Q&A system with ChromaDB and conversation memory.

Research Assistant Dev:

The assistant ingests a PDF, stores vectors in ChromaDB, answers from the document with source citations, and remembers previous questions.

S5

Tool Calling + ReAct Agents — Deep Dive

What you learn:

How an agent decides what action to take — the thought-action-observation loop.

Research Assistant Dev:

The assistant becomes a real agent with three tools: search_web, calculator, and read_document. It traces every reasoning step in the console.

Milestone
S6

Multi-Agent Systems — LangChain + CrewAI

What you learn:

How to split complex tasks across specialized agents that collaborate.

Research Assistant Dev:

The assistant becomes a team: Research Agent + Writer Agent + Reviewer Agent. Built twice — once in LangChain and once in CrewAI.

S7

LangGraph Workflows + Evaluation Basics

What you learn:

How to build stateful, conditional workflows and measure whether your AI system actually works.

Research Assistant Dev:

The assistant routes queries to the right node (coding / factual / analysis). Students measure retrieval quality and fix poor results.

Milestone
S8

FastAPI + Streamlit + Deployment

What you learn:

How to wrap an AI system in an API, build a UI, and deploy it to a public URL.

Research Assistant Dev:

The entire Research Assistant — RAG, memory, agents, routing — deployed as a Streamlit web app. Live. Shareable. Portfolio-ready.

Month 2

The Project Phase

Students work in teams of 3-5 to build and deploy a real Agentic AI product. Every team member contributes to the backend, RAG pipeline, agents, and deployment — so everyone gets full-stack AI exposure. Guided lectures and weekly doubt sessions keep teams unblocked.

Month 2 Timeline

Sessions 1-2

Project planning, GitHub setup, backend API with FastAPI

Sessions 3-4

RAG pipeline with Qdrant + autonomous agent workflows

Sessions 5-6

Multi-agent collaboration + monitoring and optimization

Sessions 7-8

Docker, deployment, final presentation and viva

Project Types Students Can Choose From

AI Chatbot / Domain Assistant
Custom intelligent assistant for a specific domain
LangChainFastAPIStreamlitChromaDB

Build a smart chatbot for a specific domain — HR assistant, legal query bot, customer support agent, or college FAQ system. Integrates prompt engineering, memory, and a clean deployed UI.

Research & RAG Intelligence System
Document Q&A with citations and semantic search
QdrantLangGraphPydanticFastAPI

An AI system that ingests multiple documents (PDFs, articles, reports) and answers questions with citations. Includes metadata filtering, source highlighting, and retrieval quality scoring.

Multi-Agent Automation Workflow
Autonomous pipeline for content, data, or analysis tasks
CrewAILangGraphLangChainDocker

A team of AI agents that autonomously executes a multi-step workflow — content generation pipeline, competitive news automation, data analysis and report writing, or code review agent.

Real-World Domain Agent
Vertical AI agent for a specific industry use case
ReAct AgentExternal APIsStreamlitFastAPI

A practical AI agent built for a real domain: e-commerce product recommender, healthcare query assistant, finance news summarizer, or an agriculture advisory agent. Includes tool calling and external API integration.

Milestone Outcome

Every Month 2 project ends with a **live deployment**, a **public GitHub repository**, and an **individual viva** where each student explains their contribution. This becomes the student's first real AI portfolio item.

Outcomes

What a Student Gains

After completing AgentX, a student is not an AI expert — and that is not the goal. They are someone who understands how Agentic AI systems are built, can build basic versions independently, and can carry their learning forward on a job or in a startup.

Skills

  • Set up a professional Python AI project from scratch
  • Build a working RAG pipeline with vector database storage
  • Design and implement ReAct agents with custom tools
  • Build multi-agent systems using LangChain and CrewAI
  • Create stateful workflows with conditional routing in LangGraph
  • Deploy AI applications using FastAPI + Streamlit + Docker
  • Debug agent failures using logs and intermediate step traces

Portfolio

  • 1 deployed AI web app from Month 1 (accessible via public URL)
  • 1 team project from Month 2 (real-world domain, live deployment)
  • Clean public GitHub with meaningful commit history
  • Architecture diagram and README for every project

Career Readiness

  • Can explain RAG, agents, and multi-agent systems in an interview
  • Knows how to position AI projects on LinkedIn and resume
  • Understands system design of an end-to-end AI application
  • Foundation to self-learn any new LLM framework or tool independently

Academic Alignment

Why Colleges Choose AgentX

Designed to slot seamlessly into academic institutions, bridging the gap between computer science theory and industrial deployment.

Closes the AI Skills Gap

Most CS curricula cover ML theory but zero practical agent development. AgentX fills what 4 years of college misses in 2 months — with hands-on, deployable output.

Students Graduate with a Portfolio

Every student completing AgentX has a live deployed project on GitHub. Recruiters notice this. Placement teams notice this.

No Infrastructure Required

AgentX uses cloud-based LLM APIs and free-tier vector databases. Students need only a laptop and internet connection. No GPU servers required.

Complements, Does Not Replace

AgentX is designed as an add-on program, not a replacement for existing curriculum. It runs on weekends and requires no changes to academic scheduling.

Industry-Aligned Stack

The tech stack — Python, LangChain, FastAPI, Docker — matches current AI engineer job descriptions. Students learn what companies are actually hiring for.

Measurable Outcomes

Every session has an assignment, every month ends with a deployed project. Progress is visible, assessable, and demonstrable to any external stakeholder.

Ready to bring AgentX to your institution or enroll your students?

Contact us to schedule a demo session, discuss batch requirements, or request a custom program proposal.