Category: Machine Learning

Business Education, Career Management, Google APIs, Machine Learning, Python

Levels of AI Agents by their Architectural Components

By Architectural Components, I mean, how the agent makes decisions. The levels describe the core components of an AI agent and how it processes information.  These include agents like Simple Reflex, Model-Based Reflex, Goal-Based, and Learning Agents. Let’s just get familiar with these before we develop agents in the future blogs.

  • Simple Reflex Agents: Respond directly to current inputs without considering past experiences or future goals. 
  • Model-Based Reflex Agents: Maintain an internal model of the environment to handle partially observable situations and make more informed decisions. 
  • Goal-Based Agents: Use their models to search for sequences of actions that achieve specific goals, providing a sense of purpose. 
  • Utility-Based Agents: Go beyond goals to choose actions that maximize a “utility function,” leading to more rational and optimal outcomes. 
  • Learning Agents: Can improve their performance over time by learning from their experiences and interactions with the environment. 
by admin
Business Education, Entrepreneurship, Fintech, Machine Learning

Should countries and incubators create affordable platforms?

In today’s economy, digital platforms are infrastructure. Just like roads and electricity powered the industrial era, platforms like Uber, ChatGPT, and Amazon power the digital age. But here’s the catch—most of these are privately owned and expensive to access.

For instance, Indian entrepreneurs, especially from Tier-2 and Tier-3 cities, platform cost is a barrier, not just a factor.

  • Startup failure rate in India is ~90% (IBM Institute).
  • API usage fees are pricing out early-stage innovators.
  • AI tools like ChatGPT are unaffordable for most students or bootstrapped founders.

But India has shown what’s possible. It has created platforms already. For instance, it has created UPI, a public payments infrastructure, open to all and ONDC is building an alternative to Amazon/Flipkart, enabling local sellers to go digital.

Why not extend this to:

  • AI (IndiaGPT?)
  • Mobility & logistics (open Uber-alternative)
  • Cloud & data infrastructure (sovereign stack)

By creating public digital platforms, the government can:

  • Lower entry barriers for startups
  • Democratize access to cutting-edge tech
  • Retain intellectual property within India
  • Fuel job creation and inclusive innovation

India has entrepreneurs. It needs platforms that empower them. Incubators should create them too.

by admin
Machine Learning, Python

Consuming Chatgpt API in Python

Here’s the Python code I used in the class. I’m storing my key in the environment variable ‘OPENAI_API_KEY.’ Ram Subramaniam.

from openai import OpenAI
import os
openai.api_key = os.getenv("OPENAI_API_KEY")
print ("\n")

client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
mess = "beyond 120, any increment in IQ does not necessarily imply higher creativity"
prompt = "restate the following statement: " + mess
print ("Original Message : "+mess)
for x in range (0, 5):
 chat_completion = client.chat.completions.create(messages=[{"role": "user", "content": prompt}, ],model="gpt-3.5-turbo")
 print("Revised Message : " +chat_completion.choices[0].message.content)
by admin
Fintech, Machine Learning, R programming, Tech

IIM Fintech Lecture – Indian Institute of Management – Bangalore

In this course I plan to cover items 1 through 5. I’m flexible with the level of participation from students in programming exercises in step 5 (ML). Depending on backgrounds and goals, students may choose not to program (why learn programming if you can hire a developer for 10$/Hr?) or actively program in R and/or Python.

0) How much ‘Fin’ and how much ‘Tech’?

1) How banks’ reluctance to innovate/adopt led to Fintech. Banks! You can’t bank on them.

2) The ultra-personalisation of financial services through technology – the driver behind Fintech

3) Case Studies – Peer-to-peer lending, Crypto-currencies, Robo-advisers, online-only digital banks

4) My Money Karma

5) Machine Learning using R (or Python)

Ram Subramaniam Stanford
by admin
Career Management, Entrepreneurship, Machine Learning, Tech

The 10 most in-demand skills of 2019, according to LinkedIn

I don’t like it, but the four of the 6 top skills required are IT skills! Please read on. Analytical reasoning is not exclusive to IT and I haven’t considered it as an IT only skill! Others are.

1) Cloud Computing
2) Artificial Intelligence
3) Analytical Reasoning
4) People Management
5) UX Design
6) Mobile Application Development

Click the URL and read on……

The 10 most in-demand skills of 2019, according to LinkedIn

Ram Subramaniam Stanford
by admin