Available for AI Engineering roles & freelance work

Building AI that
thinks. acts.
automates.

Agentic AI Developer

AI Engineer from Kathmandu specialising in autonomous tool-use agents, full-stack RAG pipelines, and LLM workflow orchestration with n8n. I build systems that reason, retrieve, and act — then deploy them with FastAPI, React & Vercel.

bibek@ai:~$ python agent.py --mode agentic
✓ Tool-use agent initialised
✓ FAISS vector index loaded (1,247 chunks)
✓ n8n workflow webhook active
bibek@ai:~$ 
Bibek Joshi
Claude APILangChainn8nFAISS

What I Build

From autonomous AI agents to production-deployed RAG systems — I build end-to-end AI solutions that work in the real world.

Agentic AI Systems

Autonomous tool-use agents with multi-step orchestration, persistent memory, and real-time reasoning using Claude & LangGraph.

RAG & Knowledge Systems

Full-stack RAG pipelines — document ingestion, FAISS vector search, semantic retrieval, and contextual LLM responses.

AI Workflow Automation

n8n-based multi-channel workflows that connect LLMs, Google Sheets, Gmail, Discord, and external APIs for business automation.

LLM API Integration

Production-ready integration of Claude, OpenAI GPT, and Gemini APIs with streaming, tool-calling, and prompt engineering.

Full-Stack AI Deployment

FastAPI backends + React/Vite frontends deployed on Render & Vercel — REST APIs, inference pipelines, scalable architecture.

Next.js Web Development

Freelance Next.js, React, and MERN Stack development — from landing pages to full production web applications.

Engineering meets Intelligence

I'm Bibek — an AI Engineer from Koteshwor, Kathmandu. My background is in civil and transportation engineering, which gave me a strong foundation in structured problem-solving and data-driven analysis.

I transitioned into AI by completing professional training in Data Science, ML, and most recently Agentic AI Engineering. Today I specialise in building systems where LLMs don't just generate text — they reason, use tools, retrieve knowledge, and execute multi-step workflows autonomously.

When I'm not building AI systems, I'm a certified yoga & meditation trainer and a part-time instructor with The Art of Living — which I find brings real clarity to complex engineering challenges.

10+
AI Projects
5+
Certifications
7+
Yrs Experience

Academic Background

M.Sc. Transportation Engineering
IOE, Tribhuvan University – Pulchowk
2015 – 2017
B.E. Civil Engineering
Kathmandu Engineering College, TU – Kalimati
2010 – 2014
Kathmandu, Nepal 🇳🇵
Open to remote AI roles worldwide

Skills & Tools

The stack I use to build production-grade AI systems.

Agentic AI & Automation
n8nLangGraphMCPAI AgentsMulti-Agent OrchestrationAgentic RAGDockerVPS
Generative AI & LLMs
Claude APIOpenAI GPTGemini APIPrompt EngineeringRAGContext-Aware AILangChainLangSmith
RAG & Vector Search
FAISSVector EmbeddingsSemantic SearchDocument ChunkingContext RetrievalBeautifulSoup
Deployment & Full-Stack
FastAPIStreamlitReactViteNext.jsNode.jsExpress.jsRenderVercel
ML & Deep Learning
Scikit-learnTensorFlowKerasCNNTransfer LearningNLPTF-IDFEmbeddings
Programming & Data
PythonSQLPandasNumPyGitMySQLMongoDBJupyter

Work History

From infrastructure engineering to AI systems.

AI Engineer (Training & Projects)
Agentic AI Engineering — Broadway Infosys
Apr 2026 – Jun 2026

Built agentic AI systems, RAG pipelines, LangChain workflows, multi-agent orchestration, and n8n automation. Projects: AI Job Tracker, Hotel Automation System, RAG Chatbot.

Web Developer (Freelance)
React · MERN Stack · Next.js · WordPress
2023 – Present

Production web apps for clients including a restaurant POS & accounting app, dynamic business sites, and e-commerce storefronts.

Freelance Engineering Analyst
Infrastructure & Data Analysis
2017 – 2024

Data-driven technical reports and structured analysis for road infrastructure projects. Strengthened analytical reasoning across field datasets.

Featured Work

Real AI systems built and deployed — not demos. Each project has a problem, a stack, and a measurable outcome.

Agentic AI · Tool-Use

AI Job Application Tracker

Problem

Manually tracking job applications across emails, calendars, and spreadsheets is fragmented and error-prone. Needed an autonomous agent that handles the entire workflow end-to-end.

Stack
Claude APIReactViteNode.jsGmail APIGoogle SheetsGoogle CalendarOAuth2NDJSON
Outcome
Agent autonomously reads emails, detects duplicates via memory, logs to Sheets, creates Calendar follow-ups, and sends Gmail confirmations with real-time NDJSON-streamed reasoning.
n8n · AI Workflows

Hotel Automation System

Problem

Hotel operations across concierge, booking, housekeeping, and revenue were siloed and manual. Needed an AI-powered, multi-workflow system to connect all of it.

Stack
n8nOpenAI GPT-4o-miniGoogle SheetsGmailDiscordWebhooks
Outcome
7 interconnected n8n workflows automate full hotel ops. AI Concierge Agent with memory handles guests; LLM chains power bookings, upselling, sentiment analysis, and revenue reporting.
Full-Stack RAG

AI RAG Chatbot System

Problem

Standard LLMs hallucinate on domain-specific content. Needed a full-stack system where users interact with their own documents through context-aware retrieval.

Stack
PythonFastAPIReactViteLangChainFAISSLLM APIsRenderVercel
Outcome
End-to-end pipeline: document ingestion → chunking → embeddings → FAISS retrieval → LLM response. Live at chat-app-rag.vercel.app.
RAG · Vector Search

Website Intelligence Agent

Problem

Websites hold valuable info that can't be queried conversationally. Needed a system to scrape any site, build a searchable vector index, and answer questions about it.

Stack
PythonLangChainFAISSGemini APIStreamlitFastAPIBeautifulSoup
Outcome
Full RAG pipeline: scrape → chunk → embed → FAISS → Gemini response. Conversational UI at ai-websiteagent-bibek.streamlit.app.
Agentic AI · Memory

Goal-Oriented Learning Assistant

Problem

Generic chatbots forget context and can't adapt to learning goals. Needed an assistant that tracks objectives and tailors every response to the user's defined goals.

Stack
PythonLangChainGemini APIStreamlitFastAPI
Outcome
Persistent memory tracks goals across sessions; dynamic prompts inject them into each response. FastAPI + Streamlit at learning-chatbot-bibek.streamlit.app.
Deep Learning · CNN

Cat vs Dog CNN Classifier

Problem

Demonstrate end-to-end deep learning deployment: fine-tune a CNN with transfer learning and serve real-time predictions through a REST API and interactive UI.

Stack
PythonTensorFlowKerasCNNTransfer LearningStreamlitFastAPI
Outcome
Fine-tuned CNN classifies images with high accuracy. FastAPI REST API handles uploads; deployed Streamlit UI for real-time predictions.

Trainings & Certs

Scroll to see all certifications →

Agentic AI Engineering Training
Broadway Infosys Pvt. Ltd.
Apr – Jun 2026
AI AgentsLangChainMulti-Agent OrchestrationAgentic RAGVector DatabasesMemory & StateMultimodal AIAI Safety
AI & Machine Learning Training
Broadway Infosys Pvt. Ltd.
Feb – Apr 2026
Deep Learning (ANN, CNN)NLPGenerative AILLMsPrompt EngineeringRAG
Data Science & Machine Learning
Broadway Infosys Pvt. Ltd.
Jun – Aug 2025
Statistics & ProbabilityMachine LearningNLP & Text MiningSQLData VisualizationStreamlit & FastAPI
Workflow Automation with n8n
Udemy
2025
AI Workflow AutomationAgent OrchestrationAPI IntegrationDocker Deployment
Python, Django & Web Development
Broadway Infosys Pvt. Ltd.
Apr 2023 – Feb 2024
PythonDjango FrameworkREST API IntegrationReact (Frontend)MERN Stack

Beyond Code

The Art of Living

Mindfulness & Service

Part-time instructor contributing to service and mindfulness initiatives.

Yoga & Meditation

Certified Trainer

Certified in yoga and meditation — practices that sharpen focus and clarity in complex engineering work.

Music

Devotional Songs

Passionate about devotional and mindfulness-oriented music as a source of inspiration.

Let's Connect

Open to AI engineering roles, freelance projects, and collaboration. Reach out — I respond fast.

bibek.joshi91@gmail.com ↗
Kathmandu, Nepal +977 9843702195 Open to remote worldwide