Engineering intelligent developer systems.
I am a |
Computer Science senior at UET Lahore. Architecting graph database project states, local RAG pipelines, and deterministic Pydantic AI systems to eliminate version control complexity.
Interactive AI Workspace
Recruiters and developers can query this portfolio dynamically. Input commands into the terminal emulator below to unlock system design files and professional insights.
Work History
Focused entirely on engineering intelligent workflows, software architectures, and mentoring academic pipelines.
AI Systems Engineering Intern
Bookme.pk
- ↳
Engineered an automated, production-grade conversational AI assistant for booking workflows.
- ↳
Pioneered the transition from LangChain to Pydantic AI to enforce strict type-safe outputs, leading to deterministic tool execution and eliminated runtime state errors.
- ↳
Architected advanced Retrieval-Augmented Generation (RAG) pipelines using Graphiti RAG to model complex multi-node user preferences and historical sessions.
- ↳
Constructed a hybrid NLP and LLM-driven parameter extraction system, increasing data point collection accuracy from 75% to 92%.
- ↳
Integrated real-time flight query tools and third-party APIs directly into the agent's decision loops, streamlining booking operations.
Teaching Assistant — Object-Oriented Programming (OOP)
UET Lahore, Computer Science Department
- ↳
Tutored and mentored a class of 40+ undergraduate students, breaking down core paradigms: inheritance, polymorphism, encapsulation, and abstraction.
- ↳
Conducted weekly lab tutorials, guided hands-on debugging sessions, and evaluated software architectural quality in student project submissions.
- ↳
Designed algorithmic challenges to cultivate standard software engineering design principles and clean code practices.
System Design Deep Dive
Detailed overview of AutoGit, a semantic version control agent transforming workspace change telemetry into active graph dependency indices.
AutoGit
AI-Powered Version Control & Workspace Graph Database
An intelligent Visual Studio Code extension that transforms version control from a manual chore into a semantic, context-aware automated workflow.
AutoGit Challenge
Traditional relational databases or text embeddings failed to capture the architectural relationships (AST nodes, imports, data flows) between modified files, resulting in generic context representations.
System design choice
Pivot from flat vector spaces to a Neo4j Graph Database. By modeling code bases as active dependency graphs, AutoGit maps modifications to exact dependency clusters, retrieving exact contextual nodes in O(1) time.
Measurable engineering impact
Drastically improves repository hygiene, eliminates context drift during branch switching, and provides high-fidelity, automated, review-ready pull request context.
Intelligent Systems Inventory
Explore secondary production-grade applications covering machine learning pipelines, multi-tenant schedules, and mobile optimization systems.
Pediatric Growth Diagnostics & AI Nutritional System
A comprehensive, intelligent mobile application designed to track infant developmental metrics and deliver clinically sound dietary guidance.
AI-Driven School Operations & Timetabling SaaS
A premium, multi-tenant learning management system (LMS) and operations portal that automates administrative and educational workflows.
Hotel Management & Resource Optimization System
A production-grade mobile app optimized for hotel administrative staff to streamline guest check-ins, payments, and live room inventories.
Technical Stack
A comprehensive index of tools and workflows I use to construct intelligent software systems.
Academic Vector
UET Lahore
Bachelor of Science in Computer Science
Period: 2022 — 2026
// AI & Intelligent Workflows
// Languages
// Backend & Databases
// Developer Tooling & Tools
Clinical Machine Learning Pipelines
Investigating explainable diagnostic parameters and data balancing pipelines for young adult health metrics.
Hypertension Risk Prediction in Young Adults
Clinical Predictors & Explainable AI (SHAP) • 2024
An advanced machine learning pipeline and academic paper studying early hypertension risk factors in young adults using genetic and biometric indicators.
- ✔ Engineered a medical ML classification pipeline comparing Random Forest, Support Vector Machines, and Gradient Boosting Trees.
- ✔ Integrated SHAP (SHapley Additive exPlanations) to provide explainable local and global feature attribution for clinical transparency.
- ✔ Achieved a 97.6% hypertension classification accuracy using a custom SMOTE balanced dataset.
Certifications & Badges
Large Language Models Specialization
H2O.ai & Coursera
- ↳ Completed rigorous learning paths covering prompt engineering, fine-tuning techniques, and hardware optimization of massive open-source models.
- ↳ Gained hands-on experience deploying open LLMs locally and evaluating pipeline metrics.
Python & SQL (Basic & Advanced)
HackerRank
- ↳ Verified deep language expertise in Python algorithmic puzzles and complex multi-join SQL query optimizations.
Get In Touch
Looking for a technical AI engineer to build scalable software, LLM workflows, or graph databases? Let\'s discuss your startup projects or engineering pipelines.