As 2025 wraps up, this year looked very different from my earlier ones. I did less physical reading overall, with most of my learning happening digitally. That said, here are my notable offline reading milestones. Deep Work by Cal Newport : Picked this up based on Reddit recommendations. A solid read on managing distractions and cultivating focus. Did it completely change my habits? I am still figuring that out. Mathematical Statistics by S.C. Gupta and V.K. Kapoor : A classic from my college days. Unexpectedly relevant again, and extremely helpful for revisiting the foundations behind my more traditional AI learning.…
-
-
Drawn using Gemini by giving the architecture design plan.
-
1. Download Ollama from https://ollama.com/download 2. To download and run the model, use the command: ollama run qwen2.5:72b 3. Once model download is completed, just type your prompts! Meanwhile, my poor GPU Next, here is a bonus. Access the model from a LangChain python script. Your Ollama service is already running at http://localhost:11434/ Install dependencies: pip install langchain-ollama Create a test_client.py program: from langchain_ollama import ChatOllama from langchain_core.prompts import ChatPromptTemplate from langchain_core.output_parsers import StrOutputParser import time # 1. Configuration # We use the specific model tag you pulled in Ollama MODEL_TAG = “qwen2.5:72b” print(f”— Connecting to Local Ollama ({MODEL_TAG}) —“)…
-
Just published my new tutorial on Medium: Full Stack Architecture Tutorial with Python, SQLModel, Strawberry, GraphQL and FastAPIRead: https://ninethsense.medium.com/one-small-full-stack-architecture-tutorial-with-python-sqlmodel-strawberry-graphql-and-fastapi-9de975c54465?sk=3c5f75201ee8486b23280f7fdc5ec67b
-
Vefify: https://academy.langchain.com/certificates/ae7mwrbmfo
-
Vefiry: https://academy.langchain.com/certificates/4vwsres05f
-
The modern software architect is often bottlenecked not by a lack of knowledge, but by the sheer volume of context they must maintain. Scalability constraints, compliance guardrails, cloud service limits, and legacy entanglements create a cognitive burden that makes pure design difficult. Read my new article on – How AI is redefining the Application Architecture Design process – https://www.linkedin.com/pulse/how-ai-redefining-system-design-process-praveen-zohtc
-
Navigating the fast-evolving world of LLMs requires more than just innovation, it demands discipline. In my latest piece, A Sanity Guide to Version Control for LLMs, I break down practical strategies to bring order, traceability, and confidence to AI development workflows. From managing model iterations to ensuring reproducibility, this guide offers a structured approach to keep experimentation aligned with enterprise needs. Read my full article here – https://www.linkedin.com/pulse/sanity-guide-version-control-llms-praveen-nair-pmp-architect-irpgc/
-
Read my latest article on agentic AI adoption in Platform Engineering on Medium. https://ninethsense1.medium.com/platform-engineering-2-0-the-rise-of-ai-native-devops-5b6c9510decc?sk=eadef43fd8cf604c21105ff055d9aead We spent years asking developers to care about infrastructure, aka follow the DevOps culture. AI-Native Platform Engineering now effectively says, “Actually, forget it. The AI handles the infrastructure now! You just focus on code”
-
This article is a write up of my experience on hosting Qwen 2.5 the 0.5B model in Raspberry Pi 2 using Llama.cpp. Qwen 2.5 is one of the small SLM with 0.5B parameters so a small development board like Raspberry Pi can hold it. RPi 2 Model B comes with 900Mhz speed and only 1GB of memory. But to be honest, setting up the project might take 1-2 hours, and the prompt execution is only some 1-2 tokens per second. So you need to be patient. Let us begin! Step 0: Pick up the Raspberry Pi 2 from the attic.…