Muhammad

Salman

AI Engineer

I build full-stack LLM applications — from RAG chatbots and internal automation tools to the APIs and frontends that run them. Primarily Python, deployed on AWS.

70%

Hosting costs cut

99%

Service uptime

10+

Projects shipped

Get In Touch

View Projects Download CV

About Me

AI Engineer who builds full-stack LLM applications — from RAG chatbots and internal automation tools to the APIs and frontends that run them. I work primarily in Python and deploy on AWS, with shipped projects that improved reliability and cut hosting costs.

Experience

AI Engineer

Dec 2025 - Present

The Insolvency Group

Bolton, UK (Remote)

Automating internal business tools using AI to reduce operational costs and inefficiencies
Built RAG-based systems and applied ML techniques to improve internal decision support and workflow accuracy

AI Engineer

June 2025 - Aug 2025

Anviro

Islamabad, Pakistan

Developed RAG-based chatbots that enhanced customer response accuracy by leveraging real-time data retrieval.
Deployed scalable AI services on AWS and Render, ensuring 99% uptime and seamless user experience.
Designed and optimized vector search pipelines using embeddings and semantic retrieval to improve chatbot relevance and reduce hallucinations.

AI/ML Intern

June 2024 - Aug 2024

Aim Lab

Islamabad, Pakistan

Led a team of 5 in developing an email template generation project using Gemini API for text generation and Stable Diffusion for images, enhancing project delivery time and quality.
Built and maintained web applications using React.js and Git for version control.
Deployed backend services on AWS EC2 and frontend on Vercel using React.js, reducing hosting costs by 70% and improving application reliability.

Education

B.S. Data Science

Aug 2022 - Present

National University of Computer and Emerging Sciences

(FAST NUCES)

3.3

/ 4.00 CGPA

Activities:

• Dean's List — Fall 2025• Team Operations at FAST Data Science Society (22-23)• Team Data Quest at Nascon'23• Team PR/Marketing at FAST Data Science Society (23-24)• Team Technical Workshops at Google Developers Group on Campus (24-25)

Key Achievement

ICPC Asia Topi Regional Contest

Ranked 134th out of 2000+ teams in the regional programming contest

Asia Topi Online Programming Contest • Nov 2024

Data Structures

Object-Oriented Programming

Certifications

LangChain for LLM Application Development

DeepLearning.AI•Dec 2024

AI Frameworks

View

Excel Business Intelligence: Power Pivot, DAX and Data Modeling

LinkedIn•Aug 2024

Business Intelligence

View

Excel PivotTables: Mastering PivotTables and PivotCharts

LinkedIn•Aug 2024

Data Analysis

View

Supervised Machine Learning: Regression and Classification

DeepLearning.AI•Jun 2024

Machine Learning

View

Introduction to AWS

Datacamp•Jun 2024

Cloud Computing

View

Harnessing the Power of Data with Power BI

Microsoft•Mar 2024

Visualization

View

Preparing Data for Analysis with Microsoft Excel

Microsoft•Feb 2024

Data Preparation

View

Pandas

Kaggle•Jun 2023

Data Science

View

Research

Improving Robustness and Interpretability of Deep Learning–Based Intrusion Detection in Software-Defined Networks

Conditionally Accepted

ICIT 2026•Paper ID: 172

Proposes an enhanced deep learning–based intrusion detection system for Software-Defined Networks, focusing on improving model robustness against adversarial inputs and interpretability via explainability techniques.

View Paper

Featured Projects

Explore my portfolio of data science and AI projects, showcasing innovative solutions and cutting-edge technologies.

AI/ML

MailFusion

React.js

Gemini API

Unsplash

A React-based application that streamlines email template creation and management. Leverages AI text generation (Gemini) and image generation (Unsplash) for dynamic, customizable templates.

Code

Data Science

Electric Load Forecasting

Python

React

scikit-learn

pandas

Predicts future electricity demand using advanced preprocessing, clustering, and forecasting models. Features an interactive React dashboard for visualization and data-driven energy management.

Code

AI/ML

Cold Email Generator using LangChain

Python

LangChain

Streamlit

Chromadb

Automates personalized cold email creation based on job descriptions. Extracts requirements, matches them with portfolio links, and crafts contextually relevant emails.

Code

Data Engineering

Metro Near Real-Time Data Warehouse

Java

SQL

MySQL

Eclipse

Implements a near real-time data warehouse using Java for ETL and SQL for database setup and OLAP analysis. Features master and transactional data loading, star schema design, and advanced analytics.

Code

Visualization

TB-Dashboard

HTML

CSS

JavaScript

D3.js

An interactive data visualization platform featuring timelines, treemaps, and maps. Offers sorting, filtering, and highlighting for deeper insights into TB-related data.

Code

AI/ML

Music Recommendation System

Python

MongoDB

Flask

TensorFlow

Identifies similar songs based on extracted audio features. Integrates lyrics extraction via speech recognition and uses MongoDB for data storage.

Code

AI/ML

MailAssist

Next.js

TypeScript

Supabase

OpenAI API

Shopify API

Full-stack AI customer support platform built on a RAG pipeline. Handles ticket triage, generates on-brand responses using vector retrieval, integrates Shopify for live order data, and includes an analytics dashboard for tracking resolution rates and response times.

Code Live Site

AI/ML

Siamese Legal Clause Similarity

Python

PyTorch

BiLSTM

CNN

Siamese network architectures (BiLSTM and CNN with Attention) for detecting semantic similarity between legal clauses. Trained on 295,000+ clause pairs across 395 categories. Achieved 99.98% accuracy without pre-trained transformers.

Code

AI/ML

GNN Patient Triage

Python

PyTorch Geometric

GraphSAGE

Reproduces a Nature Scientific Reports paper using GraphSAGE for emergency patient triage classification. Includes cosine similarity graph construction, SMOTE balancing, and rigorous train/test leakage prevention.

Code

Data Engineering

MLOps Pipeline (Airflow + MLflow)

Python

Apache Airflow

MLflow

scikit-learn

End-to-end MLOps pipeline with Airflow DAGs for orchestration and MLflow for experiment tracking and model registry. Includes parallel task execution, branching logic, and automated model registration based on accuracy thresholds.

Code

Want to see more of my work?

View All Projects on GitHub