Dayyan Saeed

AI & ML Engineer specializing in Autonomous Multi-Agent Systems and RAG Pipelines.

Based in: Lahore, PakistanCurrently: AI & ML Engineer, SolutyicsExperience: 1 year

Dayyan is an experienced AI & ML Engineer with a strong background in developing LLM-powered applications and architecting complex multi-agent systems. He excels at building end-to-end predictive models and scalable RAG engines using modern frameworks like LangChain, FastAPI, and Pinecone.

Experience

AI & ML Engineer

SolutyicsOn-site, Lahorework2026 – Present

Summary

Engineered and optimized end-to-end predictive models, driving the full machine learning lifecycle from data preprocessing and feature engineering to deployment. Developed and integrated LLM-powered applications leveraging prompt engineering, memory management, and advanced reasoning workflows. Architected autonomous multi-agent systems with tool-calling, recursive reasoning, and dynamic state management. Built Retrieval-Augmented Generation (RAG) pipelines using vector databases to enable semantic search and document intelligence.

What I did

Architected multi-agent workflows where specialized agents handled research, reasoning, validation, and synthesis.
Designed orchestration, failure handling, and stopping conditions for agentic systems.
Implemented recursive reasoning for task decomposition and verification.

Results

Reduced redundant computation and improved execution speed on parallelizable tasks by optimizing multi-agent workflows.
Enhanced output consistency through the implementation of validation steps and structured handoffs between agents.

Machine LearningLLMPrompt EngineeringMulti-agent systemsRAGVector databases

Skills

Technical

Python

Prompt Engineering

RAG

LangChain

LangGraph

OpenAI API

Gemini API

Agentic AI

NumPy

SQL

OpenAI Codex

Tavily

Claude Code

CopilotKit

Docker

Git

Scikit-Learn

PostgreSQL

Pandas

AWS (EC2, S3)

C++

TypeScript

TensorFlow

MySQL

Firebase Firestore

DSPy

TruLens

GitHub Actions

Firebase

BeautifulSoup

Matplotlib

NLTK

OpenCode

SQLite

Lexical boosting

Multi-tenant architecture

Celery

Schema Mapping

LoRA

React

Next.js

Fine-tuning

Upsert Operations

Projects

GitHub Intelligence MCP

Backend Developerpersonal

Built an MCP server providing AI agents with structured GitHub repository intelligence including metadata, issues, pull requests, commits, contributors, releases, and repository search. Implemented asynchronous GitHub REST API integration with Pydantic validation, pagination, retries, rate-limit handling, structured errors, automated testing, Docker, and CI quality gates.

PythonMCPPydantichttpxDockerGitHub REST APICI

Luxury Watch Extractor & Catalog Management System

AI Engineerpersonal

Built an AI-powered web data extraction platform using Tavily and Gemini to discover, extract, structure, and validate luxury watch information from web sources. Developed Flask APIs, relational database workflows, and Celery background jobs for automated catalog ingestion, upsert operations, and scheduled data extraction.

PythonTavilyGeminiFlaskSQLCeleryPostgreSQLSQLite

RAG Engine

Full Stack Developerpersonal

Built a full-stack Retrieval-Augmented Generation (RAG) application for document ingestion, semantic search, and citation-based question answering using LangChain, embeddings, and Pinecone. Implemented FastAPI APIs with JWT authentication, PostgreSQL persistence, Redis rate limiting, multi-tenant data isolation, and Server-Sent Events (SSE) streaming; deployed the frontend on Vercel.

PythonFastAPILangChainPineconePostgreSQLNext.jsJWTRedisSSEVercelVoyage AI embeddings

Generative UI Agent Platform

AI Engineerpersonal

Developed a Generative UI application using CopilotKit and LLM agents for dynamic user interactions and context-aware UI generation. Built agent-driven workflows with LangGraph and backend tool execution to support conversational task completion and dynamic data-driven interfaces.

PythonCopilotKitLangGraphLLMsReact

Tasdeeq

personal

A project focused on OCR for printed Urdu/Nastaliq invoices using a fine-tuned vision-language model for structured field extraction.

Qwen2-VL-2BLoRAQaariEasyOCR

Education

Bachelor of Science in Computer Science

University of Central PunjabComputer ScienceOct 2021 – Oct 2025

Recognition

Certifications

100 Days of Machine Learning & Deep Learning

CampusX

Generative AI Engineering

CampusX

Agentic AI & Multi-Agent Workflows

CampusX

AWS Academy Cloud Foundations

AWS Academy

AI For Everyone

DeepLearning.AI

Hire DayyanReach out about a role, a contract or a conversation.For recruiters

Dayyan's twin is AI, it can make mistakes.

More people on Kanvis

Browse more people Create your own Kanvis