Rafiou Diallo

Machine Learning & Cybersecurity

Applied AI · CyberSecurity · Interpretable ML

Machine Learning & Cybersecurity Enthusiaste specializing in Applied AI, Security, and Interpretable ML. I build intelligent systems at the intersection of machine learning, cybersecurity, and real-world deployment with a focus on robustness, interpretability, and scale.

GitHubLinkedIn
Scroll

Featured Projects

Production-grade ML and security systems built with a focus on interpretability, robustness, and real-world deployment.

NIST Cybersecurity Q&A System (RAG)

Retrieval-Augmented Generation system delivering grounded cybersecurity guidance from official NIST publications.

Query
FAISS
Context
Mistral
Answer

Query flows through FAISS retrieval to Mistral LLM with automatic source attribution

  • Built end-to-end RAG pipeline using Mistral-7B-Instruct, LangChain, and FAISS
  • Implemented 4-bit quantization for memory-efficient inference (~4GB VRAM)
  • Engineered recursive PDF scraping and validation for NIST SP 1800 documents
  • Generated semantic embeddings with Sentence Transformers for sub-100ms retrieval
  • Enabled automatic source attribution for every answer
PythonLangChainFAISSMistralNLPWeb Scraping
View on GitHub

Malware Image Classification using CNNs

Interpretable deep learning system for malware family classification from raw binaries.

High attention
Structure

Model attention highlights structural fingerprints used for classification

  • Converted malware binaries into 128x128 RGB images
  • Designed custom CNNs and fine-tuned ResNet50, achieving 98.4% test accuracy
  • Optimized architectures with Optuna
  • Implemented Grad-CAM for per-family interpretability
  • Built misclassification analysis pipeline for edge-case discovery
PyTorchCNNsResNet50OptunaGrad-CAM
View on GitHub

ML Network Intrusion Detection (CICIDS2017)

Production-scale ML system detecting real-world network attacks across millions of flows.

Flows
Features
LightGBM
SHAP
Protocol
Duration

SHAP reveals protocol-level attack signatures across 2.2M+ flows

  • Processed 2.2M+ network flows across 10 attack categories
  • Achieved 98% macro-F1, 99% recall on Bot attacks
  • Addressed extreme class imbalance with targeted resampling
  • Optimized LightGBM across 40+ configurations
  • Applied SHAP to reveal protocol-level attack signatures
PythonLightGBMSHAPPandasNetwork Security
View on GitHub

Semi-Supervised Anomaly Detection in System Logs

Sequence-based anomaly detection system for large-scale distributed logs.

p95 threshold
Block sequence
Anomaly
Normal

Long-tail anomalies prioritized for investigation, not hard classification

  • Built LSTM next-event prediction model using PyTorch
  • Learned normal execution patterns from unlabeled data
  • Scaled analysis to millions of HDFS log lines
  • Aggregated losses using p95 / max / mean for block-level detection
  • Emphasized exploratory analysis over brittle classification
PyTorchNLPLSTMPandasNumPy
View on GitHub

Experience

Professional experience in security operations, full-stack development, and enterprise system engineering.

Carleton University

Information Security Co-op

Sep 2025 – Present·Ottawa, ON
  • Conducted 50+ vulnerability assessments using Nessus & Qualys
  • Automated compromised-credential intelligence parsing with Python
  • Deployed phishing simulation to 30,000+ users
  • Investigated DNS/SPF misconfigurations and supported WAF tuning
  • Authored vulnerability management SOPs and governance documentation
NessusQualysDefenderPythonPowerShellTCP/IPWAF

myAIpathway.org

Software Developer Intern

Feb 2025 – Sep 2025·Remote
  • Built full-stack applications using React, Node.js, Python
  • Contributed to FRED, a platform reducing food waste
  • Worked in agile environment with Git-based workflows
ReactNode.jsPythonGitAgile

Société Générale

Software & Security Intern

May 2024 – Jul 2024·Guinea
  • Built production banking systems used in live operations
  • Automated document workflows, reducing paper usage by 50%
  • Implemented dashboards, audit logs, and REST APIs
  • Supported network security operations and mentoring
ReactNode.jsREST APIsSQLNetwork Security

Skills & Tooling

Technical expertise across machine learning, cybersecurity, and software engineering.

Machine Learning & AI

  • PyTorch, Scikit-Learn
  • Deep Learning (CNNs, LSTM, Transformers)
  • Natural Language Processing (NLP)
  • Retrieval-Augmented Generation (RAG)
  • Model Interpretability (SHAP, Grad-CAM)
  • Hyperparameter Optimization (Optuna)
  • Representation Learning & Embeddings
  • Semi-supervised & Unsupervised Learning

Data & Analytics

  • Pandas, NumPy, Matplotlib
  • Data Preprocessing & Feature Engineering
  • Exploratory Data Analysis (EDA)
  • Large-scale dataset handling
  • Class imbalance handling & resampling
  • Statistical aggregation & optimization

Cybersecurity & GRC

  • Vulnerability Management
  • Network Security & Intrusion Detection
  • Governance, Risk & Compliance (GRC)
  • WAF Operations & Traffic Analysis
  • DNS / SPF / Email Security
  • Threat Modeling & Security Awareness
  • Zero-day & anomaly-based detection

Software Engineering

  • Python, JavaScript, SQL
  • React, Node.js, FastAPI
  • REST APIs & Backend Systems
  • Automation & Scripting (Python, PowerShell)
  • Git, Agile Development
  • System Design & Documentation
  • Performance Optimization

Security & Dev Tooling

  • Tenable Nessus, Qualys
  • Microsoft Defender
  • FAISS, LangChain
  • Jira, Bitwarden
  • Linux / CLI workflows

Education

Carleton University

B.C.S. Honors Computer Science

Minor in Mathematics

GPA: 3.5

Expected Graduation: 2027

Let's Connect

Open to opportunities in ML, cybersecurity, and applied AI systems.

EmailGitHubLinkedIn