Niroshan Ranapathi

Agentic Workflow Automation (GenAI, CrewAI, n8n, Flowise)

Designed and deployed end-to-end agentic automation systems using GenAI tools

including LangGraph, n8n, and Flowise, enabling intelligent task delegation,

context retention, and tool integration across workflows. ( AWS Bedrock, Anthropic

Claude, Llama 2, Mistral )

Developed reusable agent templates for lead generation, web research, outreach,

and multi-tool orchestration with persistent memory.

Conversational AI & Virtual Assistant Development

Built full-stack chat applications powered by LLMs for use cases including

customer support, task execution, and dynamic prompt routing.

Implemented real-time streaming, speech-to-text, and tool-triggering via

structured prompt systems, ensuring scalability and fast user interactions.

Generative AI Ecosystem Knowledge – Deep understanding of AI orchestration

frameworks (LangChain, LlamaIndex, Haystack) and cloud-based AI offerings

(AWS Bedrock, Vertex AI, Azure Machine Learning).

LLM Proficiency – Covering prompt engineering, fine-tuning, and integration into

real-world applications. Skilled in working with multiple LLM providers (OpenAI

GPT/O families, Anthropic Claude, Google Gemini, xAI Grok) and open-source

models (LLaMA, DeepSeek, Mistral).

Graph Database Expertise – In-depth knowledge of graph database concepts, data

modeling, and query languages (e.g., Cypher), with practical experience using

Neo4j, Amazon Neptune

Government Commercial Cloud integration integrated application development for

Singapore Building Construction Authority-related applications (CorenetX and Project

Portal).

Manage cloud infrastructure and perform VAPT in accordance with government

security guidelines (IM8, CISCAT, OWSAP , ITGC)

Building LLM-enabled Chatbot Applications for Financial Analysis and Cybersecurity

Compliance Management (GPT 4o, claude, Lamma)

Building LLM-enabled multi-agent system development and NLP voice-enabled chatbot

development

POC and research related to RAG integration (VectorDB, Pre-Retrieval, Post-Retrieval,

MLOps practices, MLFlow)

Deploy machine learning models and evaluate model performance.

Developed applications for Microsoft Windows OS, including client Software

Development Kits (SDKs), MS DirectML, WinML, and MS Graph.

Establish the required tests for continuous integration and deployment, and implement

and oversee the automation pipeline

ML-enabled application deployment with Terraform, Docker, and serverless

Building a highly available application for the Singapore defense ministry (OneNs,

SmartICT)

Conducted security assessments and developed cyberproof applications

Troubleshooting and fixing bugs in existing microservices

Setting up a data pipeline and automating infrastructure with cloud formation and

terraform