Usama's Lab — Bridging Academia & Industry

Forging AI-powered

AI & Full Stack Engineer and researcher translating applied AI into production impact at Lean Automation. Focused on RAG systems, LLM fine-tuning, predictive maintenance, and medical imaging research.

terminal://usama-lab

  ██╗   ██╗██╗      █████╗ ██████╗ 
  ██║   ██║██║     ██╔══██╗██╔══██╗
  ██║   ██║██║     ███████║██████╔╝
  ██║   ██║██║     ██╔══██║██╔══██╗
  ╚██████╔╝███████╗██║  ██║██████╔╝
   ╚═════╝ ╚══════╝╚═╝  ╚═╝╚═════╝ 
        L A B  ·  A I / M L        
[00]sysinitializing industrial-ai stack...
[01]ragFAISS retriever online for plant incident logs
[02]llmdomain adaptation complete: Llama 3 + Grok
[03]iotkafka stream stable at 10,000+ events/sec
[04]ml predictive maintenance model accuracy > 90%
[05]resDHNN MRI classifier validated at 93.2%
[06]api12+ microservices healthy · sub-50ms event latency
[07]ok all systems nominal ✓
$run --profile cv-sync --target usama-lablive
v1.0.0
Dec. 2025

Timeline

Professional Journey

Experience across AI engineering, full-stack product development, and applied medical AI research.

AI & Full Stack Engineer

Feb 2024 – Present
Lean AutomationPakistan

Architecting industrial AI systems spanning RAG, LLM fine-tuning, predictive maintenance, and event-driven microservices at production scale.

  • Built RAG systems with FAISS + LLMs for complex industrial log intelligence
  • Fine-tuned Llama 3 and Grok for domain-specific industrial reasoning
  • Designed Kafka pipelines processing 10,000+ sensor events/second with sub-50ms latency
  • Decomposed monolith into 12+ microservices handling millions of daily events
PythonTypeScriptLlama 3GrokFAISSKafkaNode.jsNestJSDocker

Full Stack Engineer

Jan 2023 – Jan 2024
Lean AutomationPakistan

Built and shipped production web platforms and automation pipelines across analytics, APIs, CI/CD, and deployment workflows.

  • Delivered 5+ production full-stack apps with 99.5% availability SLAs
  • Implemented GitHub Actions + Docker CI/CD for zero-downtime deployments
  • Automated reporting via SQL procedures and APIs, reducing manual effort by 80%
  • Built real-time dashboards for KPI and anomaly monitoring
ReactNext.jsNode.jsNestJSSQLDockerGitHub Actions

Research Assistant

Mar 2024 – Feb 2025
COMSATS University Islamabad (Abbottabad Campus)Abbottabad, Pakistan

Led thesis research on medical AI for brain tumor analysis, focusing on model design, dataset engineering, and publication-ready outputs.

  • Developed a Dual-Head Neural Network (DHNN) for tumor type + grade prediction
  • Built end-to-end MRI preprocessing pipelines with OpenCV and Pandas
  • Achieved 93.2% classification accuracy with EfficientNetB0 backbone
  • Co-authored and submitted work to Elsevier Results in Engineering
PythonTensorFlowKerasOpenCVPandasMedical Imaging

Bachelor of Science in Computer Science (BSCS)

Mar 2021 – Jan 2025
COMSATS University Islamabad — Abbottabad CampusPakistan

CGPA: 3.56 / 4.0. Focused on AI/ML systems, software engineering, and data-driven problem solving.

  • Thesis: Multi-Class Brain Tumor Classification and Grade Estimation using DHNN
  • Coursework: AI, ML, DSA, DBMS, Operating Systems, Distributed Systems
  • Built research and production projects spanning healthcare and industrial AI
AIMachine LearningComputer VisionDistributed SystemsOOPDatabases
Journey continues...

Portfolio

Featured Projects

Research-backed production systems delivering measurable impact across healthcare and industrial sectors.

Featured
Research
2024

Multi Class Brain Tumor Classification and Grade Estimation using Dual Head Neural Network

Dual-Head Neural Network leveraging EfficientNetB0 for multi-class tumor classification achieving 93.2% accuracy with extreme computational efficiency (4.0M parameters, 15ms inference).

93.2% accuracy15ms inference4.0M parameters
TensorFlowKerasMedical AIComputer VisionResearch
Featured
Research
2024

A VAE LSTM Architecture for Early Fault Signatures, Time to Failure Projection, and Root Cause Pathway Analysis

VAE-LSTM architecture for industrial predictive maintenance with 48-hour advance fault detection. Achieved $2.3M annual savings with 20% downtime reduction.

92% accuracy48h advance warning$2.3M savings
PyTorchFastAPIKubernetesKafkaIndustrial AI
Featured
2024

HarmFilter App

Multi-model BiLSTM hate speech detection and prevention tool addressing harmful online content in South Asian contexts. Actively educates users to promote respectful online communities.

Multi-dataset trainingSouth Asian contextReal-time detection
BiLSTMNLPMachine LearningSocial ImpactMulti-Model
Featured
2024

FarmGuardian

AI-based decision support system for Punjab farmers. Provides disease detection and yield forecasting through Urdu-based interface, helping minimize crop losses and optimize resources.

Real-time disease detectionYield forecastingFarmer-friendly UI
Deep LearningMobile AppAgricultureUrdu NLPSocial Impact
Featured
2024

Eventure

Innovative tourist platform revolutionizing travel in Pakistan with personalized journeys, real-time security alerts, and seamless vendor integration for a dynamic travel experience.

Real-time alertsLocal vendor integrationCustomized planning
Next.jsReactTourismSecurityMultilingual

Work in Progress

Workbench

Active experiments and prototypes. Things that are being built, broken, and rebuilt.

~/Usamarana01/active
$

industrial-rag-core

Retrieval-Augmented Generation stack for industrial logs using FAISS + LLM pipelines for failure triage and root-cause assistance.

95%
May 2026
$

predictive-maintenance-suite

VAE-LSTM and ANN/RNN pipelines for oil & gas fault forecasting with production-grade monitoring.

88%
Apr 2026
$

iot-event-mesh

Kafka-based sensor ingestion mesh sustaining 10,000+ events/second with sub-50ms latency and zero data loss objectives.

92%
May 2026
$

dhnn-brain-tumor-study

Dual-Head Neural Network research workflow for simultaneous tumor type and grade estimation from MRI scans.

90%
Mar 2026
git status --all