KMKashindra Mahato

Kathmandu, Nepal · Open to remote

Kashindra Mahato
Full-Stack AI Engineer

4+ years shipping production AI — SEC filings parsed into SDG alignment reports, financial documents turned into structured audit outputs, agentic RAG replacing workflows that used to take analyst teams. Across FinTech, CPG, and Security.

Explore

Where I work best

Four years shipping AI in regulated, high-stakes domains — where wrong answers cost money, not just accuracy points.

01

FinTech & Risk

Asset allocation, credit risk modeling, and automated financial document analysis.

02

ESG & Sustainability

UN SDG alignment research, SEC filing analysis, and company sustainability scoring using agentic RAG.

03

Cyber Security

Intelligent threat detection and behavioral anomaly monitoring systems.

04

Document Intelligence

End-to-end extraction and classification pipelines for unstructured financial and legal documents.

Featured Projects

Deep dives into production-grade AI systems designed for scale and performance.

View all projects
NLP

Agentic RAG System

Production agentic RAG pipeline with LangGraph orchestration, LangChain tooling, Azure AI Search, and SSE streaming. Deployed custom LLMs on Azure with semantic search via Universal Sentence Encoder.

NLPLangGraphAzure AI Search
View on GitHub
ESG

UN SDG Intelligence Suite

End-to-end platform for analyzing companies' UN SDG alignment using SEC EDGAR filings, hybrid RAG (semantic + BM25 + ChromaDB), a LangGraph multi-phase pipeline, and a full-stack web platform with HITL review.

LangGraphChromaDBClaudeSEC EDGAR
Finance

Optimal Assets Allocation

Portfolio optimization system predicting optimal asset weights to minimize risk and maximize return. Combines classical statistical methods with modern ML for the NEPSE market.

FinancePredictive AnalysisPython
View on GitHub

Working on something hard?

If you're pulling data from messy sources, building something agentic, or trying to make a language model actually reliable in production — that's the kind of work I do.