Author

DhanushKumar

@danushidk507

articles: 18
total claps: 231
top claps: 51
avg read: 11.3 min
active: 2024-08-07 → 2026-07-19

Wiki topics

LLM · Large Language Models · 17 AGT · AI Agents · 7 AI · AI · General · 5 EVAL · Evaluation & Benchmarks · 4 EDU · Education & Learning · 4 FT · Fine-tuning & Adaptation · 4 RAG · RAG & Retrieval · 3 ML · Machine Learning · 3

Top tags

#llm ·6 #agentic-ai ·4 #rags ·3 #claude-code ·3 #gemma ·2 #genai ·2 #ai-agent ·2 #anthropic-claude ·2 #calling-function ·1 #google ·1

SPT

DH DhanushKumar · Stackademic

FunctionGemma — Function Calling model by Google

FunctionGemma is a new lightweight Gemma 3 (270M-parameter) model from Google tuned specifically for function (tool) calling.It acts as an…

Dec 23, 2025 · 5 min read · 51 claps #gemma #calling-function #google

FunctionGemma — Function Calling model by Google

SPT AI TCH

DH DhanushKumar · Artificial Intelligence in Plain English

Customizing PaliGemma: A Guide to Fine Tuning for Targeted Applications

For Code : https://github.com/Idk507/Finetuning_Paligemma/tree/main

Aug 07, 2024 · 15 min read · 39 claps #paligemma #multi-model #llm

AI HUM

DH DhanushKumar · Artificial Intelligence in Plain English

Securing Generative AI: Risks, Threats, and Mitigations

Generative AI (GenAI) systems — powered by large language models (LLMs) and deep learning — offer transformative capabilities but…

Nov 04, 2025 · 16 min read · 33 claps #llm #security #llm-poisoning

Securing Generative AI: Risks, Threats, and Mitigations

AI SOC

DH DhanushKumar · Towards AI

Agent Swarming

Agent swarming is a multi-agent AI pattern in which several specialized agents collaborate on the same problem through a shared…

May 09, 2026 · 9 min read · 30 claps #agent-swarm #agentic-ai #genai

AI TCH

DH DhanushKumar · Stackademic

Ollama launch with Claude Code , Codex,Droid, Opencode

Ollama exposes a local/cloud model host that implements an Anthropic-compatible messages API; this makes it straightforward to run open or…

Jan 29, 2026 · 11 min read · 22 claps #ollama #claude-code #xcode

Ollama launch with Claude Code , Codex,Droid, Opencode

DH DhanushKumar · Towards AI

Cog-RAG: Cognitive-Inspired Dual-Hypergraph RAG

Retrieval Augumented Generation boosts LLM by grounding them in external knowledge .Traditional RAG simply retrieves flat text chunk which…

May 10, 2026 · 9 min read · 11 claps #cograg #rags #graphrag

Cog-RAG: Cognitive-Inspired Dual-Hypergraph RAG

DH DhanushKumar

Harness in AI Agents

A harness in AI agents is the runtime and control plane that turns a raw model into something that can actually do work. In Anthropic’s…

May 19, 2026 · 12 min read · 10 claps #ai-agent #harness-engineering #openai

HUM AI

DH DhanushKumar · Stackademic

Manifold-Constrained Hyper-Connections (mHC) — Breaking down !!

Before diving into mHC , lets revise about Residual and Hyper connections

Feb 08, 2026 · 21 min read · 9 claps #mhc #resnet #hyper-connection

Manifold-Constrained Hyper-Connections (mHC) — Breaking down !!

DH DhanushKumar · Artificial Intelligence in Plain English

Agentic AI -IV : A Deep Dive into the Evaluator-Optimizer Workflow and GAIA Benchmark

In the rapidly evolving landscape of artificial intelligence, ensuring the accuracy and reliability of AI-generated outputs is paramount…

May 15, 2025 · 4 min read · 9 claps #agentic-ai #lanchain #gaia

Agentic AI -IV : A Deep Dive into the Evaluator-Optimizer Workflow and GAIA Benchmark

HUM AI

DH DhanushKumar

RAFT: Teaching LLMs to Read, Not Just Retrieve

Why your RAG pipeline keeps hallucinating — and how Retrieval Augmented Fine-Tuning fixes it from the inside out.

Jun 18, 2026 · 6 min read · 6 claps #rags #sftp #raft

RAFT: Teaching LLMs to Read, Not Just Retrieve

AI ART

DH DhanushKumar · Artificial Intelligence in Plain English

AI Agent Orchestration Patterns

Agent architectures exist on a spectrum of complexity, and each level introduces coordination overhead, latency, and cost. Use the lowest…

May 12, 2026 · 15 min read · 3 claps #agentic-ai #orchestration #microsoft-agent-framework

AI SPT DSN

DH DhanushKumar · GoPenAI

Beyond the LLM Call: Managing Prompts, Caching, and Routing in FastAPI

When building Generative AI applications, it’s incredibly easy to get a prototype up and running by throwing a few hardcoded prompts into…

Jul 14, 2026 · 12 min read · 3 claps #prompt-engineering #fastapi #python-development

Beyond the LLM Call: Managing Prompts, Caching, and Routing in FastAPI

DH DhanushKumar · GoPenAI

Evaluating RAG Systems with Mean Reciprocal Rank (MRR)

Retrieval-Augmented Generation (RAG) has become a cornerstone of modern Generative AI systems. Instead of relying solely on the internal…

Sep 08, 2025 · 4 min read · 3 claps #mrr #rr #llm-evaluation

Evaluating RAG Systems with Mean Reciprocal Rank (MRR)

HUM AI

DH DhanushKumar

Is One Layer Enough?

“Sometimes the biggest breakthrough in machine learning doesn’t come from adding more parameters — it comes from realizing that we may not…

Jul 19, 2026 · 18 min read · 1 claps #lrs #llm #attention

DH DhanushKumar · Stackademic

Agent Hooks (Middleware) in LangChain

LangChain’s agent middleware (also called hooks) let developers intercept and customize an AI agent’s behavior at key points in its…

Feb 15, 2026 · 8 min read · 1 claps #agent-hooks #middleware #ai-agent

AI TCH

DH DhanushKumar

I Gave an AI the Ability to Write and Run Its Own Code.

The most dangerous program is one that runs without anyone watching.” — every senior engineer who has debugged production at 2 a.m.

May 25, 2026 · 24 min read · 0 claps #sandbox #coding-agents #autonomous-agent

I Gave an AI the Ability to Write and Run Its Own Code.

AI ECO

DH DhanushKumar · Artificial Intelligence in Plain English

Claude Agent Management

Claude Agent Management is not just a single agent feature and Claude Managed Agents for a pre-built hosted agent runtime that handles…

May 03, 2026 · 8 min read · 0 claps #claude #anthropic-claude #agent-management

AI SPR

DH DhanushKumar

DINOv3: Self-Supervised Vision Model by Meta AI

DINOv3 is a self-supervised vision foundation model that learns rich, general-purpose visual representations without needing any labeled…

Jun 08, 2026 · 6 min read · 0 claps #meta #dinov3 #vit