Home About Experience Ventures Projects Skills Contact
// hello, world!

Ved Prakash
Pathak

Open to Work

AI|ML System Architect specializing in document intelligence, multimodal AI platforms, and production-grade LLM systems. Building AI-first products that extract value from unstructured data at scale.

Ved Prakash Pathak
scroll

Who I Am

AI|ML System Architect & Builder

Currently VP at Legend's Crafts (DocExtract.ai), leading AI product strategy for an intelligent document processing platform — extracting structured data from PDFs, invoices, and scanned images at scale.

I specialize in LLMs, RAG pipelines, Computer Vision, and industrial AI — turning unstructured data into production-grade intelligence. Open to impactful roles and collaborations.

0
Years in AI/ML
0
Projects Built
0
Certifications
0
Products Shipped
Download Resume
Python90%
LLM / RAG Systems88%
Document Intelligence / OCR90%
Computer Vision85%
TensorFlow / PyTorch85%
Cloud Platforms (AWS / GCP / Azure)75%
FastAPI / Backend85%
MLOps / Docker / Kubernetes72%
NLP87%

Where I've Worked

Legend's Crafts · DocExtract.ai Mar 2025 – Present

Vice President

  • Leading product strategy and AI development for DocExtract.ai, an AI-powered document intelligence platform extracting structured data from PDFs, scanned images, and invoices in Excel, CSV, and JSON formats.
  • Architected an intelligent document processing pipeline integrating OCR, layout analysis, and LLM-based field extraction enabling seamless ERP and CRM integrations.
  • Overseeing end-to-end platform development including invoice parsing, automated data entry elimination, custom field mapping, batch processing, and audit trails for compliance.
  • Driving platform adoption and business strategy for SMEs, finance teams, and operations workflows requiring high-accuracy document automation at scale.
Legend's Crafts · DocExtract.ai Jul 2024 – Mar 2025

AI Developer

  • Designed and built Intellifind.ai, a cutting-edge multimodal AI platform empowering SMBs to manage and interact with business data across spreadsheets, images, videos, and PDFs — without any technical expertise.
  • Developed customizable AI workspaces for Knowledge Database, Inventory management, and Invoice processing with real-time data interaction via a conversational AI chatbot.
  • Implemented RAG-based retrieval pipeline for document Q&A, enabling automated inventory updates and CSV/JSON invoice conversion with multi-format file support.
Self-Employed Apr 2023 – Present

Freelance AI Developer & Machine Learning Engineer

  • Built a sub-millimetre precision Computer Vision system capable of detecting hairline cracks as small as 0.05mm on 1mm rivets — enabling automated pass/fail quality control on high-throughput industrial conveyor lines.
  • Engineered a real-time wire harness measurement system using ArUco marker-based spatial calibration and morphological skeletonization to measure wire lengths and detect branch junctions with high accuracy, replacing manual inspection entirely.

Independent Ventures

📦

RactoGateway

Python Library · Open Source

Authored and published RactoGateway, a Python library designed to streamline AI/ML API gateway patterns, simplifying routing, authentication, and request management for AI-powered backends.

📧

EM-telligence.ai

Founder · In Active Development

Offline Email Intelligence Platform leveraging locally hosted LLMs for privacy-preserving email analysis, smart reply generation, and workflow automation — without cloud dependency. Integrates Ollama-hosted models (Qwen, LLaMA variants) with GGUF quantization for efficient local inference.

Building in stealth

What I've Built

Skills & Tools

🐍

Languages

PythonJavaScriptHTML & CSSSQL
🤖

AI Domains

NLPComputer VisionGenerative AIDocument IntelligenceIndustrial AI
🧠

AI Techniques

RAGNERPOS TaggingText ClassificationOCRGAN TrainingSentiment AnalysisSemantic Extraction
🔧

ML Frameworks

TensorFlowPyTorchScikit-learnXGBoostKeras
☁️

Cloud & DevOps

AWSGCPAzureDockerKubernetesVertex AIMLflow
🦙

LLM Tooling

Ollamallama.cppGGUF QuantBitsAndBytesQwenLLaMABERTGPT

Backend & APIs

FastAPIFlaskStreamlitWebSocketRESTful API
📊

Data Science

NumPyPandasMatplotlibSeabornPlotlyETLApache Spark
📝

Text Representation

n-gramsTF-IDFWord2VecGloVeBERTGPT Embeddings

Education

🎓

Bachelor of Engineering — Mechanical Engineering

Rajiv Gandhi Proudyogik Vishwavidyalaya, Bhopal

2015 – 2019 CGPA: 7.54

Certifications

Google Cloud & Machine Learning certifications — View Google Developer Profile →

Computer Vision

Computer Vision

Visual data analysis and interpretation on GCP.

ML Fundamentals

ML Fundamentals

Foundational machine learning principles and techniques.

ML Pipelines

ML Pipelines

Designing and deploying efficient ML workflows.

Production ML

Production ML

Deploying ML models into production environments.

Recommendation Systems

Recommendation Systems

Designing personalized recommendation algorithms.

TensorFlow

TensorFlow

Building and deploying deep learning models.

MLOps

MLOps

End-to-end ML workflows with Airflow, Kubeflow, Vertex AI.

🏅

+More on Google Cloud

View all badges on the Google Developer Profile.

Get In Touch

Open to opportunities

Currently VP at DocExtract.ai and actively open to new opportunities in AI/ML engineering, product leadership, and system architecture. Let's build something impactful together.

Greater Noida, U.P., India