Prujith AI, Generative AI, Agentic AI, Python
No reviews yet

AI Systems Architect & GenAI Specialist
3+ years of technical leadership spanning to cutting-edge Agentic AI from RPA (Automation Anywhere). Expert in deconstructing high-level abstractions (LangChain, CrewAI, AutoGen) into high-performance PocketFlow implementations. Specialized in building JSON & Code Agents with secure execution, grounded in Finite State Machines, Universal Approximation Theorem and first-principles mathematics.

Proven track record in implementing PyTorch-based generative models (VAEs/GANs) and production-grade RAG/ReAct patterns. Beyond engineering, I serve as a Generative AI Instructor, mastering the full lifecycle from API Testing (Postman) and Web Scraping to Vector DB orchestration and SQL backend design with Serialization languages of XML, YAML.

Teaching Philosophy:
I believe in fostering curiosity, encouraging innovation, and building a strong foundation of technical concepts. My teaching approach is focused on hands-on learning, real-world applications, and collaborative problem-solving to prepare students for practical challenges.

Teaching Style and Methodology:
Interactive Learning: Engaging students through discussions, Q&A sessions, and brainstorming activities.
Project-Based Approach: Assigning real-world projects to enhance practical knowledge and problem-solving skills.
Step-by-Step Guidance: Breaking down complex concepts into simpler modules for better comprehension.
Customized Learning Plans: Tailoring lessons based on individual learning styles and pace.
Technology-Driven Tools: Leveraging AI tools, coding platforms, and simulations for enhanced learning experiences.

Results and Achievements:
Assisted teams in automating processes using RPA tools, reducing manual effort and improving efficiency.
Delivered AI-powered solutions by implementing Machine Learning and Deep Learning techniques.
Trained peers and colleagues on automation frameworks, scripting, and AI-driven approaches, receiving positive feedback.
Developed reusable solutions, saving time and costs in project implementations.

What Makes Me a Great Teacher:
Proven ability to explain technical concepts in an easy-to-understand manner.
Passion for mentoring and inspiring students to achieve their goals.
Focused on practical applications to make learning engaging and impactful.
Continuous learner, staying updated with the latest trends in AI and Data Science to deliver cutting-edge knowledge.

Let's Connect:
Ready to inspire the next generation of AI and automation enthusiasts. Connect with me to explore how we can shape the future together!

Subjects

Python Beginner-Expert
Tamil Beginner-Expert
Artificial Intelligence Beginner-Expert
Generative AI Beginner-Expert
Agentic AI Beginner-Expert

Experience

Generative AI + RPA + Agentic AI Teacher (May, 2025–Present) at SURE ProEd

• Agentic AI with JSON Agents (structured tool calling) and Code Agents (executable code block generation), along with custom secure code execution
• Decoded abstractions of LangChain, LangGraph, CrewAI, SmolAgent, AutoGen, and Pydantic using PocketFlow.
• Finite State Machines (FSM), Universal Approximation Theorem
• Design patterns of Semantic Search, Retrieval Augmented Generation (RAG), and ReAct patterns (Observation -> Reasoning -> Action), etc
• PyTorch to implement Latent Spaces, Autoencoders, and Variational Autoencoders (VAE), Generative Adversarial Networks (GAN)
Coding Tutor - Python, Scratch Game Development (May, 2025–Present) at BeGalileo.com

• Fundamental to Advanced Python concepts range from data structures, OOP, Inheritance, Exception handling, etc.
• Custom as well as inbuilt libraries like Pandas, Numpy, Matplotlib, etc, for Data Analysis, Visualization, Mathematical Operations, etc
Prompt Engineer (Aug, 2023–Dec, 2023) at Servicepack.ai

Technologies/Skills applied: Python, LLM, Data Science, Prompt Engineering, Data analysis, Langchain, Postman, JSON, OpenAI API, Mistral API
• Created more than 1200 datasets for transcribing call recordings within the healthcare insurance sector, featuring interactions between insurance agents and customers using the GPT-3.5 Turbo model by OpenAI to orchestrate the conversational prompts.
• Achieved 95 % accuracy rate in utilizing the Mistral prompt format to enable the model to evaluate and score the performance of the agents in the dataset across various dimensions, such as Upselling, Empathy, etc.
• Employed API calls, particularly through the use of HTTP POST requests by the Postman platform iteratively till the prompts are refined, adjusted and tailored to deliver the most precise insights from the transcriptions.
• Developed various prompt engineering solutions and name a few are Sentiment Analyzer, Total Webpage Summarizer, QA and dictionary bot, Content Creator, YouTube video title generator and description based on video type/topic and keywords mentioned by the user, etc.
• One example is Custom domain conversational chat tool which I developed using Langchain and OpenAI Large Language Model (LLM) for automated information retrieval. This tool takes a URL as input, makes API requests, scrapes web content using Beautiful Soup, tokenizes text, and embeds it with OpenAI Embeddings. Utilized FAISS for vector database management. Successfully delivered highly accurate and efficient results for user queries.
Tool Room Engineer (Jun, 2018–Jun, 2019) at Titan

Tool Room Engineer

Education

Mechanical & Automation Engineering (Jun, 2014–Apr, 2018) from MAHENDRA ENGINEERING COLLEGE

Fee details

₹50–500/hour (US$0.53–5.26/hour)

Courses offered

Generative AI
- USDFREE
- Duration: 11 Days
- Delivery mode: Flexible as per the student
- Group size: Individual
- Instruction language: English, Tamil
- Certificate provided: No
Required amount of Python, Machine Learning, Deep Learning for Generative AI.
• PyTorch to implement Latent Spaces, Autoencoders, and Variational Autoencoders (VAE), Generative Adversarial Networks (GAN)
• Finite State Machines (FSM), Universal Approximation Theorem
• Agentic AI with JSON Agents (structured tool calling) and Code Agents (executable code block generation), along with custom secure code execution
• Decoded abstractions of LangChain, LangGraph, CrewAI, SmolAgent, AutoGen, and Pydantic using PocketFlow.
• Design patterns of Semantic Search, Retrieval Augmented Generation (RAG), and ReAct patterns (Observation -> Reasoning -> Action), etc
Agentic AI
- USDFREE
- Duration: 22 Days
- Delivery mode: Flexible as per the student
- Group size: Individual
- Instruction language: English, Tamil
- Certificate provided: No
Required amount of Python, Machine Learning, Deep Learning for Generative AI.
• Agentic AI with JSON Agents (structured tool calling) and Code Agents (executable code block generation), along with custom secure code execution
• Decoded abstractions of LangChain, LangGraph, CrewAI, SmolAgent, AutoGen, and Pydantic using PocketFlow.
• Finite State Machines (FSM), Universal Approximation Theorem
• Design patterns of Semantic Search, Retrieval Augmented Generation (RAG), and ReAct patterns (Observation -> Reasoning -> Action), etc
• PyTorch to implement Latent Spaces, Autoencoders, and Variational Autoencoders (VAE), Generative Adversarial Networks (GAN)