- 2025Inter-Agent Communication (A2A) - Inter-Agent Communication (A2A) facilitates the collaboration of distinct AI agents, potentially constructed with diverse frameworks, allowing them to seamlessly coordinate, delegate tasks, and exchange information
- Knowledge Retrieval (RAG) - The Knowledge Retrieval (RAG) pattern significantly enhances the capabilities of Large Language Models (LLMs) by granting them access to external knowledge bases before generating a response. Instead of relying solely on their internal, pre-trained knowledge.
- Human-in-the-Loop - The Human-in-the-Loop (HITL) pattern integrates human judgment and creativity with AI’s computational power to ensure ethical, safe, and effective operation in high-stakes environments. By involving experts in oversight and feedback loops, this approach transforms AI into a partner that augments human capabilities to produce robust and accurate results.
- Exception Handling and Recovery - To function dependably in unpredictable real-world environments, AI agents must utilize the Exception Handling and Recovery pattern to identify issues like malformed outputs or API failures. By implementing structured strategies for error handling and system recovery—such as retries, fallbacks, or human escalation—agents maintain their operational integrity and build long-term user trust
- Goal Setting and Monitoring - This pattern enbales agents to pursue specific objectives and track their progress toward achieving those goals. The core idea is to transform simple AI systems into proactive, self-managing entities capable of reliable, autonomous operation by providing them with a clear sense of direction and the means to assess their own success.
- Memory Management - We delve into the critical concept of memory management for intelligent agents, emphasizing its necessity for coherent interactions and learning.Distinguishing between short-term memory, which holds immediate conversational context and long-term memory, serving as a persistent knowledge repository.
- Multi-Agent Collaboration - Instead of a single agent attempting to handle all aspects of a challenging task, this pattern breaks down a large problem into smaller, distinct sub-problems, each assigned to an agent with the specific tools and expertise needed.
- Model Context Protocol (MCP) - A standardized interface that allows LLMs to interact with external resources like databases, software tools, and various applications, extending their capabilities beyond mere text generation.
- Learning and Adaptation in Agents - Agents can learn and adapt to become more effective, moving beyond rigid programming to autonomously improve through experience.
- Planning in Agents - Planning pattern highlights how intelligent agents move beyond simple reactions to strategically achieve complex goals. It details how agents formulate a sequence of actions from an initial state to a desired outcome
- Tool use in Agents - Tool use enables an agent to interact with external APIs, databases, servises, or even execute code, It allows the LLM decide when and how to usee a specific external function based on the user's request.
- Reflection in Agents - Reflection pattern involves an agent evaluating its own work, output, or internal state and using that evaluation to improve its performance or refine its response.
- Parallelization in Agents. - Parallelization enables agentic tasks to run multiple sub-tasks simultaneously instead of sequentially, complementing Prompt Chaining and Routing.
- Routing in Agents - Routing enables agent dynamically evaluates specific criteria to select from a set of possible subsequent actions.
- Prompt Chaining - Prompt chaining (Pipeline pattern) breaks complex tasks into smaller steps, where each LLM output feeds into the next for easier problem-solving.
- Secretary Problem - Decision Making: The Secret Behind the Secretary Problem.
- Parameter-Efficient Fine Tuning(PEFT) - A guide to Parameter-Efficient Fine Tuning.
- Poethepoet: Python task execution tool - A tool to manage and execute all the CLI commands in your project.
- Poetry: Dependency and virtual environment manager - A tool for dependency management in Python.
- Pyenv: A python version management - A python version management tool that lets you manage multiple Python versions between projects.
- 2024Linux Commands Reference Guide - A comprehensive guide to essential Linux commands. Mastering these commands is crucial for efficient server management, script writing, and troubleshooting.
- 2022SQL Functions - A guide to SQL functions
- 2021Detecting Keratoconus from Corneal Imaging using Deep Learning.
- 2020Installing Cassandra on Mac OS X - A guide to installing Cassandra on Mac OS X
- 2019Paper Summary: U-Net - Convolutional Networks for Biomedical Image Segmentation
- Paper Summary: ResNet - Deep Residual Learning for image recognition
- 2018Deep Learning; Personal Notes Part 1 Lesson 4. - Structured learning, NLP, Collaborative filtering. Dropout, Embeddings, Back Prop Through Time
- Deep Learning; Personal Notes Part 1 Lesson 3: CNN theory. - Convolutional filters, Max pooling, activations, softmax, sigmoid & submitting results to Kaggle.
- Deep Learning; Personal Notes Part 1 Lesson 2. - Learning rate, Data Augmentation, Annealing, Test Time Augmentation.
- Deep Learning; Personal Notes Part 1 Lesson 1. - Image Classification.
- Predicting FIFA World Cup 2018 using Machine Learning.
- 2017Predicting House Prices Using Linear Regression.