Summary of "Generative AI & Agentic AI with Python @ 9:30 AM (IST) by Mr.Shiva Rama Krishna Day-1"
Overview
This session was an introductory/demo for a full course on Generative AI (GenAI) and Agentic AI with Python, led by Shiva (Sham Rama Krishna). The trainer explained what GenAI and agentic AI are, why they’re trending, the course structure, teaching methods, and placement/certification support.
Key emphases:
- GenAI is evolving rapidly and demands sequential learning and continuous practice.
- There are no shortcuts (copy/paste). Practical, end-to-end project experience is essential to get jobs.
Main ideas and concepts
- AI definition: replicating aspects of human learning and decision‑making in machines (examples: learning math, remembering lecture points).
- Generative AI: creating new content (text, images, audio, video) from learned data (examples: drawing a dog from a mental model, summarizing product reviews or movie plots).
- Chatbots and GPTs: GPT = Generative Pre‑trained Transformer; chat interfaces (e.g., ChatGPT) maintain conversational context.
- Models & algorithms: recommended learning path is sequential — RNN → LSTM → encoder/decoder → transformer → GPT/LLMs. Understanding earlier algorithms helps understand newer ones.
- Agentic AI: building software agents that automate sequences of actions. Common concepts: LangChain, RAG (retrieval‑augmented generation), chains of tools.
- Practical note: continual learning and demonstrable projects are more valuable than attendance or certificates alone.
Course content and learning path (detailed)
The course is full‑stack AI/ML (end‑to‑end projects/pipelines, not just web UI development). Main topics:
- Core Python for ML: focused syntax and OOP relevant to ML.
- Data handling: NumPy, pandas, data cleaning (compulsory).
- Deep Learning: neural networks and model training.
- Computer Vision: image/video/audio handling and related models.
- Natural Language Processing (NLP): text processing and classical NLP tasks.
- Machine Learning fundamentals: regression, classification, clustering, forecasting, recommendation systems (~20 ML algorithms).
- Transformers & LLMs: Hugging Face transformers and internals of transformers.
- GenAI tools & ecosystems: OpenAI APIs, Anthropic/Claude, other LLM providers.
- LangChain & related tooling: LangChain, LangGraph/LangSmith (or similar), RAG, language integration patterns.
- Agentic AI: building single and multi‑agent flows; tools and models for agent behavior.
- Cloud deployment: demonstrations on Azure and AWS.
Projects
- Minimum commitment: 25+ end‑to‑end projects (final commitment during the talk).
- Typical breakdown mentioned:
- AI projects (deep learning/CV/NLP): several (originally stated as 7).
- Transformers/gen projects: ~8.
- Agentic projects: ~5.
- Projects are end‑to‑end: data source → preprocessing → modeling → evaluation → deployment.
Teaching methodology, assessments, and materials
- Live coding: trainer types and explains code line‑by‑line; students are expected to follow and code themselves.
- Sequential, pipeline‑centered teaching: topics build on previous topics; skipping earlier ones hinders understanding.
- Hands‑on practice: students must implement projects themselves; simply copying code is discouraged.
- Tests and practice:
- At least ~5 timed tests during the course (may go up to 10).
- Multiple mock interviews and live interview simulations.
- Final week dedicated to interviews and resume building.
- Trainer will share ~1,000+ interview questions and ~200 Microsoft certification practice questions.
- Materials & access:
- PPTs, datasets, project notes, and recordings (recordings only for paid students) shared via Google Drive with lifetime access.
- Slide/code files are shared but intentionally non‑downloadable (trainer policy to encourage practice).
- Recordings available only to students who pay extra (recordings provided from week 2 if paid).
Mentorship and doubt resolution
- Daily doubt time and chat/mentor links for questions.
- Q&A/unmute time toward session ends.
- Trainer monitors student engagement and is expected to identify who is practicing.
Placement, certification, and guarantees
- Job support:
- Resume building and placement assistance (sample resumes and job links provided).
- Emphasis on demonstrating capability via projects and validated certifications.
- Certification guidance:
- Focus on Microsoft Azure AI certifications: AI‑900 (AI Fundamentals) and AI‑102 (Azure AI Engineer Associate).
- Trainer provides exam preparation material and question dumps.
- Trainer claimed to refund the exam fee (4,000 INR cited) if students follow his dumps and still fail (subject to his conditions).
-
Hiring guarantee (trainer commitment):
If a student develops at least 10 projects on their own (out of the 25), clears a relevant certification (Microsoft/AWS), and still does not get a job, the trainer will hire the student in his own company (SS Data Solutions).
-
Market perspective: employability was described as projects (70% weight) + skills (20%) + formal certifications (10%).
Logistics, fees, prerequisites, and policies
- Duration & schedule:
- Course length: 4.5 months.
- Regular classes: Monday–Friday, 9:30–11:00 AM IST (demo week slightly different).
- Saturdays: ML classes added free of cost.
- Fees:
- INR 20,000 without recordings.
- INR 25,000 with recordings (available from week 2 if paid).
- First week (demo) is free.
- Materials access:
- Google Drive lifetime access to course materials; slide/code files non‑downloadable.
- Prerequisites:
- No formal academic prerequisites; required: interest, consistent attendance, and practice.
- Basic arithmetic and logical reasoning sufficient (10th‑grade math level).
- Language: classes in English; Hindi queries may need subtitles.
- System requirements:
- No special hardware required; a standard laptop is sufficient.
- Policy positions:
- Trainer discourages irregular attendees. Recordings are offered for irregular students; live re‑teaching for absentees is not done.
Practical advice and repeated themes
- Follow the ordered curriculum; jumping ahead prevents comprehension.
- Practice by typing and debugging code; running downloaded scripts is ineffective.
- Continuous learning is mandatory — tools and models change rapidly.
- Build and document projects; list them concisely on resumes and be ready to explain them.
- Exam‑earned certifications (Microsoft/AWS) hold value; mere participation certificates have limited market weight.
- Interview behavior: prepare a crisp, job‑relevant introduction; confidence comes from subject mastery.
Examples and analogies used
- Drawing a dog from imagination = generative creation from learned concepts.
- Amazon product reviews summarization = text summarization application.
- Movie summarization = condensing long content into a concise narrative.
- Human brain learning examples to explain sequential learning, memory, and relevance filtering.
- Learning progression reiterated: RNN → LSTM → encoder/transformer → GPT.
Tools, platforms, and models mentioned
- Libraries & frameworks: Python, NumPy, pandas, Hugging Face transformers, LangChain.
- LLM providers & models: OpenAI, Anthropic (Claude), OpenRouter, local LLMs (e.g., small 270M models for local dev).
- Cloud & devops: Azure, AWS.
- Other topics: ChatGPT, RAG, fine‑tuning, production optimization, testing/tracing (some tracing topics may be limited for agentic AI).
Specific commitments from the trainer
- Cover every item listed in the official course syllabus (syllabus/PPT to be shared).
- Deliver a minimum of 25 end‑to‑end projects and at least 5 tests (actual coverage may be higher).
- Share certification dumps and help prepare for Microsoft/Azure AI exams.
- Provide conditional hiring in his startup if student meets the 10‑project + certification requirement and still cannot find work.
Speakers and sources
- Main trainer:
- Shiva Rama Krishna (Sham Krishna, SRK) — trainer and founder of SS Data Solutions; stated 14 years of experience; affiliations mentioned: BITS, IIT Madras, IIT Raipur.
- Participant names heard in the session (questions / mentions):
- Yoges, Hari Krishna, Rafrid, Ravindra, Mohammad (Rafil), Swati, Say Basha/Basha/Bachara, Siddhi, Anoj, and other unnamed online/offline students.
- Organizations and tools referenced:
- Narish IT / NarIT, SS Data Solutions, BITS, IIT Madras, IIT Raipur, Microsoft (AI‑900, AI‑102), Amazon, OpenAI, Anthropic/Claude, Hugging Face, LangChain/LangGraph/LangSmith, Azure, AWS, OpenRouter, WAMA, ChatGPT.
Key takeaways
- The course is hands‑on and project‑heavy, emphasizing sequential mastery from basics to advanced topics (Python → ML → DL → transformers → GenAI → Agentic AI → deployment).
- Success requires consistent attendance, active practice (typing/debugging code), building independent projects (minimum recommended: 10), and obtaining at least one exam‑based certification.
- Trainer provides substantial support (materials, tests, mock interviews) and a conditional hiring promise if strict performance requirements are met.
Category
Educational
Share this summary
Is the summary off?
If you think the summary is inaccurate, you can reprocess it with the latest model.
Preparing reprocess...