Engineering Reliable AI Systems & Resilient Architecture
Hi, I'm Ved Patel. An Honours Computer Science student focusing on RAG pipeline engineering, multi-agent validation, and distributed edge systems.
How My AI Systems Are Built
Click through each pipeline node to inspect real architectural decisions, data transformations, and validation safeguards.
1. Ingestion & Chunking
Deconstructs unstructured PDF/text documents into overlapping semantic chunks to maintain local context boundaries.
Featured Technical Projects
End-to-end systems engineered with strict attention to data contracts, vector search precision, and multi-agent verification.
Smart Document Assistant
Architected a dense vector Retrieval-Augmented Generation (RAG) assistant using LangChain and ChromaDB. Utilizes IBM watsonx.ai granite-3-2-8b-instruct LLM paired with slate-125m-english embedding model to perform context-grounded Q&A with zero hallucination escape.
- •Unstructured Document Chunking & Semantic Tokenization
- •Dense Embedding Generation via IBM Slate-125m
Polyglot Tutor
Engineered a multilingual education platform powered by a Python Flask REST API and an automated Multi-Agent Verification Pipeline. Implemented a dual-agent Teacher-TA workflow where a primary instructor agent drafts learning modules and a secondary LangChain auditor validates grade-level accuracy before client transmission.
- •Client Request via Modular React Interface
- •Flask RESTful Route Dispatch & Parameter Sanitization
Edge-Delivered Systems Portfolio
Engineered a high-performance, statically exported Next.js 16 and React 19 application coupled with Cloudflare Pages Functions for serverless contact intake, strict Content Security Policies (CSP), and zero-layout-shift design tokens.
- •Next.js 16 Turbopack Static Export (output: export)
- •Cloudflare Edge Network Global Caching & Sub-50ms TTFB
Education & Experience
A rigorous academic foundation paired with competitive hackathon architecture sprints.
Honours Bachelor of Science in Computer Science
Ontario Tech University- •4-time President's / Dean's List recipient for continuous academic excellence across all completed semesters.
- •Rigorous coursework in Data Structures, Design & Analysis of Algorithms, Database Systems, Agile Methodologies, and Software Engineering Principles.
- •Active participant in departmental engineering initiatives, hackathons, and technical projects.
System Architect & Backend Engineer — HackHive 2026
Ontario Tech University CS Club- •Architected the backend and AI orchestration layer for Polyglot Tutor, a multilingual educational platform built in 48 hours.
- •Engineered an automated Multi-Agent Verification Pipeline using LangChain to enforce grade-level validation before response delivery.
- •Utilized Abstract Base Classes in Python to decouple language learning prompts from underlying model APIs (OpenAI and Gemini).
Skills & Technology Stack
Languages, vector tools, model orchestration frameworks, and system environments used across production projects.
Languages
Core programming languages used across systems, backend, and data pipelines.
AI & Generative Engineering
Orchestration, RAG architectures, model interfaces, and deep learning libraries.
Systems, Tools & Infrastructure
Development environments, containerization, edge deployment, and version control.
Web & API Frameworks
Modern frameworks for building high-performance services and web interfaces.
About Ved Patel
Dedicated to engineering robust machine intelligence systems with rigorous academic foundations.
I am an Honours Computer Science student at Ontario Tech University. Across four consecutive academic semesters, I have maintained a 3.97 GPA and Dean's List standing while prioritizing practical, hands-on systems and AI engineering.
My engineering focus centers on Generative AI Systems. I design and implement Retrieval-Augmented Generation (RAG) architectures, working with dense vector indexing in ChromaDB, semantic embeddings, and LLM orchestration with LangChain to build context-grounded tools with strict provenance.
Beyond retrieval architectures, I explore Multi-Agent Verification Pipelines—such as the automated Teacher-TA validation loop built during HackHive 2026—to autonomously audit language models for factual and grammatical integrity before client delivery.
Get In Touch
Open to discussing AI systems engineering opportunities, co-op/internship roles, and technical collaborations.
Send Direct Message
Routed through Cloudflare Pages Functions edge API endpoint.