Skip to content
Iftekhar Rafi

Projects

Applied ML, financial software, and systems work across the portfolio.

The clearest signal is end-to-end systems work: Fyndel's AI-powered financial intelligence platform, applied ML projects, data pipelines, APIs, dashboards, embedded constraints, and the implementation details that make a system trustworthy.

Preview of Fyndel Construction Financial Intelligence Platform
Applied ML and Product EngineeringNov 2025 - PresentActive company build

Fyndel Construction Financial Intelligence Platform

Co-Founder and Product Engineer

AI-powered construction finance platform turning WIP reports, cost data, workbook imports, and accounting signals into forecasts, risk signals, and operational insight.

  • Built an 800+ file strict TypeScript product surface spanning auth, dashboards, WIP reporting, entity workflows, integrations, imports, risk, and forecasting.
  • Designed RLS-backed organization isolation, role-based permissions, audit-friendly financial schemas, and workbook intelligence pipelines for customer, project, billing, cost, and accounting evidence.
Next.js 16TypeScriptSupabasePostgreSQLDrizzle ORM
Read project detailsVisit fyndel.ai
Preview of Milton Training Needs Analysis Platform
Machine LearningMulti-month development cycleProduction-ready system

Milton Training Needs Analysis Platform

ML Engineer and Full-Stack Developer

Hybrid NLP and classification system for extracting KSAs and scoring training needs from technical documents.

  • Reached 95% accuracy in KSA extraction from technical documents.
  • Reduced analysis time from weeks to minutes through automated extraction and scoring.
PythonFastAPIDockerPyTorchspaCy
Read project details
Preview of Aircraft Health Monitoring System with Machine Learning
Machine LearningExtended multi-phase developmentOpen-source contribution

Aircraft Health Monitoring System with Machine Learning

Machine Learning Engineer

Aviation anomaly detection system combining transformer modeling, isolation forest, and full-stack monitoring dashboards.

  • Reported 95%+ anomaly detection accuracy across diverse flight conditions.
  • Designed for real-time processing across 100+ sensor channels.
PythonDjangoDjango REST FrameworkPyTorchscikit-learn
Read project details
Preview of Cybersecurity Network Intrusion Detection Using Machine Learning
Machine LearningComprehensive multi-phase projectCompleted research project

Cybersecurity Network Intrusion Detection Using Machine Learning

Machine Learning Engineer

Large-scale intrusion detection workflow for brute-force attacks across 1M+ network records and 80+ traffic features.

  • Achieved 99.99% detection accuracy with both Random Forest and DNN models.
  • Reduced manual monitoring requirements by more than 80% through automated classification.
Pythonscikit-learnKerasTensorFlowPandas
Read project details
Preview of LORIS Satellite Flight Software
Embedded SystemsOct 2020 - PresentOngoing

LORIS Satellite Flight Software

Core Software Developer

Embedded C and FreeRTOS flight software work for a student-built CubeSat mission.

  • Improved module unit testing efficiency by 8% through automated test harnesses.
  • Raised pipeline efficiency by more than 10% through automation and integration work.
CFreeRTOSEmbedded SystemsSPII2C
Read project details
Preview of Subsea Infrastructure Sensory Pod
Embedded SystemsJan 2022 - Apr 2022Completed

Subsea Infrastructure Sensory Pod

Electrical/Software Developer

Microcontroller-based monitoring pod for subsea fault and contamination detection.

  • Delivered a functional prototype with integrated fault and contamination sensing.
  • Produced a 40-page technical design report that earned a grade of A.
ATmega328PEmbedded CMicrochip StudioSolidWorks3D Printing
Read project details