Malthi S S

The program is structured as a six-week intensive, comprising four weeks of core learning followed by a two-week capstone project.

1. Foundations & The AI PM Mindset (2 Hours)

Understand the AI landscape and the unique role of an AI PM

1. The AI Spectrum: From Rules > Classic ML > Deep Learning > Generative AI
2. First-Hand Tools: Rapid ideation and prototyping with ChatGPT, Claude, and Gemini
3. Hands-On Lab (0.5 hour): ChatGPT/Claude/Gemini (prompt iteration), NotebookLM (knowledge-base prototyping), Lovable (UI mockups with AI), Co-pilot (Code Gen)

2. The Engine Room: Understanding ML & LLMs (6 Hours)

Demystify core technologies to communicate effectively with engineers.

1. ML 101 for PMs: Supervised/Unsupervised, Training/Inference, Basic Metrics (Accuracy, Precision, Recall)
2. LLMs & GenAI Deep Dive: Tokens, Context Windows, Hallucinations
3. Prompt Engineering as a PM Skill: Systematic prompt design (COT, Few-Shot), testing, and prompt versioning.
4. The "LLM Fitment" Framework: Is your problem a good fit for an LLM? When to use classification vs. generation.
5. Hands-On Lab (1 hr): Use Google Colab Notebook and the OpenAI API to build a simple text classifier and a text generator. Compare outputs and discuss trade-offs.

Resources: Google Colab, OpenAI Playground, Google’s Responsible AI Practices & Prompt Engineering Guide

3. Building with GenAI: RAG (4 Hours)

Architect solutions using the most powerful GenAI pattern - RAG

1. Problem Frameworks for AI: Applying Jobs-To-Be-Done (JTBD) to AI. The "AI Canvas."
2. RAG Deconstructed: Why it's crucial. Embeddings, Vector Databases (Pinecone, Weaviate), chunking strategies.
3. RAG Architecture & Use Cases: Search, Q&A, Knowledge Assistants.
4. Security First: Data leakage prevention, access controls, PII scrubbing.
5. Use Cases: Internal wikis vs. customer-facing chatbots (with failure post-mortems).

Resources: Build a RAG pipeline with LlamaIndex + ChromaDB (open-source) vs. Databricks Vector Search (managed).

4. Agentic AI Systems (4 Hours)

Build/Automate relevant workflows via Agents

1. Agent Types: Simple (single-task) vs. Multi-agent systems (e.g., research agents)
2. Agentic AI Explained: The "Reasoning Engine." Planning, Tool Use, Memory.
3. Critical PM Questions: When to use agents? (Complex workflows > simple Q&A). Human-in-the-loop thresholds: When to escalate to humans.
4. Orchestrate agents with LangGraph (open-source) vs. n8n (low-code workflows)
5. Use a no-code platform like n8n to build a Agentic RAG-based bot on a provided document set. Then, diagram an agent for a travel planning use case.

Tools: n8n, Cursor for Agentic application development

5. Evaluating & Observing AI Systems and ROI (4 Hours)

Define success and ensure your AI product works reliably in the real world

1. Core Topics: Metrics That Matter: Task-specific (e.g., faithfulness for RAG) vs. business metrics (cost/user). Observability Layers: Input drift, latency spikes, cost per query. Feedback Loops: User thumbs-up/down → automated retraining triggers. Ethics: Monitoring demographic skew in errors.
2. Cost Modeling: Training vs. inference costs (e.g., GPT-4o vs. self-hosted Llama 3). Hidden costs: Data labeling & annotation ROI Calculation: a. Quantitative: Cost savings (e.g., support ticket deflection rate). b. Qualitative: User satisfaction lift (NPS), brand trust.
3. Hands-On Lab (1 hr): Given a case study, learners will: 1. Fill out an AI Canvas. 2. Build a simple cost/ROI model in a shared Google Sheet.

Tools: Google Sheets for ROI modeling

6.Week 6(Sunday) - Project Presentation

Meet Your Expert Instructors for the Cohort:

We’re excited to bring you two distinguished leaders from the world of Product and AI—each with deep industry expertise and a proven track record of building, scaling, and mentoring high-impact teams and technologies.

🔹 Malthi SS

Chief Product Strategy Consultant, SparkProd Consulting Pvt. Ltd.

A product strategy powerhouse, Malthi brings decades of leadership experience from PayPal, Intuit, and SAP, advising VC-backed startups and growth-stage companies on high-stakes product decisions across AI/GenAI, B2B, and FinTech. She specializes in crafting winning product-market strategies, de-risking investment decisions, and transforming teams into engines of product excellence.

As a mentor with NASSCOM, T-Hub, and leading startup missions, she has guided 10+ ventures across their growth journeys. With 2,500+ hours of product workshops, visiting faculty roles at IIM Udaipur, Master’s Union, and Scaler, and her acclaimed podcast Product Talk with Malthi, she is celebrated as the Women Product Leader of the Year 2024.

🔹 Aniket

Senior Manager, Product – Agentic AI, Tekion Corp

Aniket brings 15 years of deep expertise in AI/ML Product Management, having led transformative initiatives at PayPal and Fidelity Investments. With nearly three years dedicated to Generative AI across pioneering startups like ASAPP and Monnai, he has shaped cutting-edge AI product strategies that solve real-world, large-scale problems. As a mentor at IIT Guwahati and Great Learning, he actively contributes to developing the next generation of AI-focused product leaders.

Target Audience and Prerequisites:

The bootcamp is designed for a specific set of professionals aiming to lead AI initiatives. The cohort is intentionally limited to a maximum of 30 participants to ensure a selective and high-quality learning environment.

Ideal Participant Profiles:

1. Product Managers: Professionals with over two years of experience who are building AI-powered features or transitioning into dedicated AI product roles.
2. Product Leaders & Directors: Individuals responsible for driving AI product strategy across multiple teams and product portfolios.
3. Business Analysts & Consultants: Professionals tasked with evaluating AI opportunities and developing business cases for their implementation.
4. Innovation Managers: Leaders spearheading AI transformation initiatives within their organizations.
5. Entrepreneurs & Founders: Individuals focused on building AI-native products and companies from the ground up

Key Take-aways:

1. A simple mental model of RAG — what “retrieval + generation” actually means and why it matters for building reliable AI products.
2. Understanding the core components — embeddings, chunking, vector stores, retrievers, and how they work together in an end-to-end flow.
3. How to frame good RAG use cases — when RAG is the right solution, when it isn’t, and how PMs can validate problem–solution fit.
4. PM’s role in RAG quality + evaluation — defining success metrics, measuring hallucinations, and designing experiments to improve accuracy.
5. Practical constraints and pitfalls — latency, cost, data quality, context limits, and common mistakes new AI PMs make (and how to avoid them).

Participants can choose to focus their projects on specific industries, including FinTech, HealthTech, EdTech, RetailTech, SaaS, E-commerce, or a custom domain.

Program Start Date - 17th Jan 2026

Duration: 6 Weeks (4 weeks core learning + 2 weeks capstone project)

Format: Online Live Classes and Virtual Office Hours

Time Commitment: 5 hours weekend classes plus 1 weekday remote office hours

Cohort Size: Limited to a maximum of 30 participants