• 🔭 I design and build end-to-end LLM applications, RAG pipelines, and ML solutions that power intelligent systems for enterprise workflows.
• 🌱 I build and optimize agentic workflows, prompt strategies, retrieval pipelines, and fine-tuning processes for production-ready GenAI systems.
• 👯 I love collaborating with other AI enthusiasts, researchers, and product builders.
• 🧠 Past projects include medical image segmentation, business-specific chatbots with RAG, document and data extraction pipelines, AI agents, OCR systems, and NLP automation tools.
• 🥅 2025 Goals: Contribute to open-source GenAI tools, grow my network, and help shape the future of applied AI.
• ⚡ Fun fact: I'm fascinated by AGI theory and enjoy translating complex AI concepts into simple stories that anyone can understand.

I've contributed to the LangChain and LangChain Community repositories, focusing on improvements to tools, utilities, and core features.

• langchain-ai/langchain-community#83
  🔢 Added AWS Bedrock Claude 4 series model cost tracking
  Enhanced bedrock_anthropic_callback.py to include token pricing data for the following Claude 4 models:

  • anthropic.claude-sonnet-4-20250514-v1:0
  • anthropic.claude-opus-4-20250514-v1:0
    Pricing based on AWS Bedrock documentation.
    Reference: AWS Bedrock Pricing
    Related Issue: #82

• langchain-ai/langchain-community#50
  🔍 Added support for additional_search_client_options in Azure Search (async)
  Extended the Azure Search async vectorstore client to accept additional_search_client_options, aligning its behavior with the sync client.
  Previously, the async client ignored this parameter and defaulted to an empty dictionary.
  Related Issue: #42

• langchain-ai/langchain#31177
  📚 Fixed outdated import paths and corrected documentation typo
  Updated the import path of DoctranPropertyExtractor from document_loaders to document_transformers to match recent package refactoring.
  Also fixed a minor typo across docs for clarity and correctness.

• langchain-ai/langchain#31107
  ⚙️ Replaced deprecated .dict() calls with .model_dump() for Pydantic v2 compatibility
  Updated multiple modules to replace deprecated .dict() usage with .model_dump() to align with Pydantic v2+ guidelines and suppress PydanticDeprecatedSince20 warnings.
  This change prepares the codebase for Pydantic v3 with no impact on functional logic.
  Related Issue: #31103

• langchain-ai/langchain-community#27
  💰 Added token usage & cost tracking for OpenAI embedding models in OpenAICallbackHandler
  Enhanced get_openai_callback() to include embedding model calls in usage and cost reporting, improving observability and enabling complete cost tracking across all OpenAI model types.
  This change aligns LangChain's callback utilities with real-world production monitoring needs.
  Related Issue: #25888

• Actively contributing additional improvements to LangChain and related projects.

I'm passionate about making AI accessible and understandable for everyone — from engineers to business leaders. I regularly write about real-world applications of AI, LLMs, RAG, and agentic workflows to share insights, simplify complex ideas, and spark meaningful conversations.

• 🔍 RAG Explained Simply for Non-Techs

Explore my work on Medium.