Mayur Kailas Shirsat

WHO I AM

Engineering the
Intelligent Edge

I'm Mayur Shirsat — an AI/ML Engineer, based in Pune, Maharashtra. I specialise in building production-grade AI systems that bridge research and reality. From fine-tuning large language models to deploying computer vision pipelines on edge hardware, I engineer solutions that don't just prototype well — they scale.

With a Master's in Data Science from Symbiosis International University and a foundation in Computer Science from Bharati Vidyapeeth, I bring rigorous academic thinking to hands-on engineering. My work spans agriculture, real-estate, healthcare, and education — always chasing the frontier of what autonomous intelligence can do.

M.Sc. Data Science & Spatial Analytics Symbiosis International University 2023 – 2025

B.Sc. Computer Science Bharati Vidyapeeth Deemed University 2017 – 2020

🛰️ ISRO AI/ML for Geodata Analysis — 2024

🐍 Python Certification — Infosys Springboard, 2023

⚡ Prompt Engineering — Great Learning

🤖 Generative AI & LLMOps GraphRAG, LoRA/QLoRA, Unsloth, LangChain, vLLM, Ollama, Prompt Engineering

👁️ Computer Vision YOLO, CNN-LSTM, Roboflow, Image Segmentation, Action Recognition, VLMs

☁️ Cloud & Deployment Azure, Docker, TensorRT, ONNX, FastAPI, n8n Agentic Workflows, Git

📊 Data & BI Power BI, Tableau, PostgreSQL, MongoDB, EDA, Statistical Modelling

Architected end-to-end vision systems using YOLO11 & YOLO26 for object detection + CNN-LSTM for action recognition and temporal behaviour modelling.
Built high-throughput OCR and visual analysis pipelines using Qwen3VL:2B, exposed via FastAPI for production inference.
Fine-tuned SLMs via LoRA on custom datasets — reduced hallucination and boosted RAG reliability in production.
Slashed GPU overhead from 95% → 15% via TensorRT acceleration + ONNX runtime optimisations — drastically cut cloud costs.
Orchestrated robust data pipelines: automated collection, EDA for class imbalance, advanced augmentation strategies.

YOLO11UltralyticsYOLO26CNN-LSTMLoRATensorRTONNXFastAPIVLM

Developed multi-agent AI systems using n8n and Vision Transformers for behavioural analytics and OCR tasks.
Built RAG-based conversational agents integrated into Streamlit interfaces with FastAPI backend model serving.
Elevated deep learning recognition accuracy by 20% via custom Roboflow pipelines and ensemble training methods.

n8nLangChainDockerVision TransformersRAGKAGStreamlitRoboflowEnsemble

Containerised and deployed AI/ML projects with Docker + Render — improved system efficiency by 30%.
Automated lead generation and SEO workflows using LLMs — drove 25% increase in acquisition and 40% higher user engagement.

DockerRenderLLMLead GenSEO Automation

Logistics and AI Automation

AI-Powered Smart Gate & Yard Management

Automated logistics ecosystem utilizing YOLOv11/v26 and CNN-LSTM frameworks for 360° damage assessment and machinery tracking. Achieved 99.9% logging accuracy by integrating VLMs for OCR and real-time Digital Twin mapping to eliminate manual yard searches.

95%Logging Accuracy

ZeroStop Throughput

360°Damage Detection

LLM & VLMDeep Learning (CNN+CNN-LSTM)Computer VisionYolo26/Yolo11FastAPIHuggingFace TransformersUltralytics

Project Info →

Education × LLM + VLM

Smart Examination System

LLM/VLM microservices — Gemma-3-4B generates exam papers, Qwen-2.5-VL extracts student answers from scanned sheets, GPT-OSS 20B auto-evaluates and scores through coordinated microservice architecture.

Gemma-3-4BQwen-2.5-VLMicroservicesVLM

View on GitHub →

Real Estate × Computer Vision

Snag Detection System

YOLO-based real-estate damage detection — wall cracks, fire damage, rusting, wire failures — with 90%+ mAP. Defect alerts trigger VLM verification and automated draft emails to builders.

90%+mAP Score

YOLOVLMWeb ScrapingAlert Pipeline

View on GitHub →

Agriculture × Multimodal AI

Multimodal LLM & CV Agricultural Assistance System

Cross-modal advisory platform synthesising text, image, and speech inputs via LLM-driven reasoning. Achieved 95% accuracy in plant disease identification. Integrated Hugging Face TTS + FastAPI for real-time market data and 5 government welfare scheme access.

95%Disease Accuracy

3Input Modalities

5Gov. Schemes Linked

LLMComputer VisionFastAPIHuggingFace TTSMultimodal

View on GitHub →

Business Intelligence

Executive Dashboard — Project Performance

Power BI dashboard backed by SQL — tracks KPIs across departments, roles, and audiences. Corporate projects identified as top revenue contributor at $510.75M. Open-sourced on GitHub.

$510MRevenue Tracked

Power BISQLEDAOpen Source

View on GitHub →

NLP × RAG × Knowledge Graphs

RAG Document Extraction System

RAG system using Meta LLaMA — extracts and summarises data from 5+ document types (PDF, DOC, CSV, links, text). Integrated NetworkX Knowledge Graph RAG. 90% accuracy, 40% better document analysis.

90%Retrieval Accuracy

40%Analysis Boost

LLaMARAGNetworkXKnowledge GraphFAISS

View on GitHub →

Full-Stack Web Application

AI-Powered Portfolio & Personal Brand Site

Responsive personal portfolio site built with semantic HTML5, modern CSS (custom properties, grid, flexbox), and vanilla JS. Features smooth scroll animations, intersection observer reveals, side-nav dot tracking, and an EmailJS-powered contact form.

100Lighthouse Score

6+Sections

0Dependencies

HTML5CSS3JavaScriptEmailJSCSS GridIntersection Observer

View on GitHub →

Data Visualization Dashboard

Business Intelligence Web Dashboard

Interactive BI dashboard built on Power BI Embedded + custom web layer. Tracks KPIs, revenue by segment, and project performance with drill-through filters and responsive layout.

$510MRevenue Tracked

Power BI EmbeddedSQLREST APIResponsive Design

View on GitHub →

AI Web Interface

RAG Document Chat — Streamlit App

Full-stack Streamlit web application enabling users to upload and chat with multi-format documents (PDF, DOC, CSV). Integrated LLaMA backend via FastAPI with real-time streaming responses and session memory.

StreamlitFastAPIPythonLLaMAREST API

View on GitHub →

Mayur Kailas Shirsat

Engineering the
Intelligent Edge

Where I've
Shipped Impact

AI / ML Engineer

AI / ML Engineer

Data Science & AI Intern

From Research
to Production

AI-Powered Smart Gate & Yard Management

Smart Examination System

Snag Detection System

Multimodal LLM & CV Agricultural Assistance System

Executive Dashboard — Project Performance

RAG Document Extraction System

AI-Powered Portfolio & Personal Brand Site

Business Intelligence Web Dashboard

RAG Document Chat — Streamlit App

Tools I
Master Daily

Let's Build
Something
Extraordinary

Mayur Kailas Shirsat

Engineering theIntelligent Edge

Where I'veShipped Impact

AI / ML Engineer

AI / ML Engineer

Data Science & AI Intern

From Researchto Production

AI-Powered Smart Gate & Yard Management

Smart Examination System

Snag Detection System

Multimodal LLM & CV Agricultural Assistance System

Executive Dashboard — Project Performance

RAG Document Extraction System

AI-Powered Portfolio & Personal Brand Site

Business Intelligence Web Dashboard

RAG Document Chat — Streamlit App

Tools IMaster Daily

Let's BuildSomethingExtraordinary

Engineering the
Intelligent Edge

Where I've
Shipped Impact

From Research
to Production

Tools I
Master Daily

Let's Build
Something
Extraordinary