Ajeet Acharya

• Develop AI solutions for accounting and tax-related classification using LangChain and OpenAI agents to handle complex decision logic.
• Design agent workflows that produce structured, schema-based outputs and manage independent tasks with a graph-based LLM framework.
• Deploy production-ready AI services with Docker and expose reusable tools via MCP servers.
• Implement proxy-based API access so OpenAI agents can securely interact with internal and external services.

• Worked with Hugging Face models locally to optimise performance and reduce dependency on external APIs.
• Fine-tuned LLMs such as LLaMA 2 and integrated Ollama and Groq for enterprise applications.
• Built LangChain agents, chains and LangGraph workflows for dynamic, multi-step AI automation.
• Implemented Redis caching for a multi-user chatbot environment, improving efficiency and scalability.
• Integrated OpenAI Realtime API and telephony systems for voice-enabled AI assistants, and used NER to extract structured insights.

• Built and fine-tuned chatbot models alongside experienced developers in Generative AI.
• Integrated Retrieval-Augmented Generation (RAG) to improve chatbot contextual accuracy.
• Implemented OpenAI APIs, function calling and Agentic RAG to enhance chatbot capabilities.

Machine Learning & AI

Generative AI & LLM Development (LangChain, OpenAI API, RAG), fine-tuning, prompt engineering, vector databases (Milvus, Chroma), chatbot development.

Backend Development

FastAPI, Flask, API development and integration.

Data & Tools

Pandas, NumPy, Scikit-learn, Matplotlib, Docker, Git & version control.

Interpersonal

Public speaking, clear communication, quick learning, composure under pressure.