Security Engineer · AI/ML Builder
I build full-stack AI and security systems — autonomous threat detection, cloud-native SaaS, and local inference — with hardened architecture at every layer.
// about
Education
Cambrian College · Sudbury
Two Graduate Certificates — Cybersecurity & AI/ML
Punjabi University · Patiala
Bachelor of Computer Applications
Cybersecurity and AI/ML graduate building at the intersection of both domains. I design and ship production systems — from autonomous threat detection platforms to full-stack AI SaaS on AWS. Security-first mindset and AI-native workflow. Based in Sudbury, ON. Open to work.
// experience
Spectrum Telecom Group Ltd.
Greater Sudbury, ON · Remote
Built and secured a full virtualized lab environment to support the company's ISO 27001 certification effort — SIEM, Active Directory, and network hardening, end to end.
Self Practitioner
Nine years of hands-on practice across custom hardware, advanced Windows administration, home networking, and AI-native development — applying academic knowledge to real-world builds.
// projects
Autonomous threat intelligence and incident response platform built to act as a Tier-1 SOC analyst — ingests logs, detects anomalies, classifies threats, maps them to MITRE ATT&CK, and generates explainable incident reports with remediation playbooks. Built with SHAP explainability for ISO 27001 / regulated-industry contexts.
Authorized local-lab security assessment agent that uses an LLM ReAct loop to probe for OWASP Top 10 vulnerabilities. Every tool call is checked against a scope.yaml allowlist before execution — LLM output is treated as untrusted, not ground truth. Maps findings to MITRE ATT&CK with CVSS v3.1 estimates and streams to a live dashboard.
AI-powered security code review tool. Analyzes submitted code across 10 languages and returns a three-part structured review: a severity-graded vulnerability report with CWE IDs and line references, a corrected version with every fix annotated inline, and a plain-language developer briefing calibrated to experience level.
Production-grade, privacy-first retrieval-augmented generation — chat with your own documents entirely on your machine. Combines hybrid search (ChromaDB vector + BM25 keyword) with automatic PII scrubbing and prompt-injection quarantine before anything is indexed. Runs local Ollama models or cloud APIs behind a Streamlit + FastAPI stack.
An intelligent mock interview platform that provides real-time feedback on spoken responses. Analyzes clarity, confidence, and structure using the STAR framework. Runs locally via Ollama for maximum privacy.
ML-driven vulnerability prioritization. Ingests CVE and exploitation intelligence from NVD, CISA KEV, EPSS, and ExploitDB, builds structured features, and trains LightGBM models to predict 30/60/90-day exploitation-likelihood windows. Surfaces results through a FastAPI backend, Plotly Dash dashboard, and Slack alerts.
// skills
// content
Documenting the journey from student to tech professional in real time — tech concepts broken down simply, honest takes on AI and the job market, and the unfiltered reality of job-hunting and immigration as a new grad in Canada. No corporate speak. No pretending to have it figured out. Follow along — it's going to be an interesting ride. 🎮💻
