// 01 — surface · DIVYANSH://VOID
security engineer
// 02 — profile
// 03 — record
1 Jun – 30 Jun 2026
May 2025 – Jun 2025
Jan 2024 – Mar 2024
// 04 — systems
End-to-end platforms with responsive, validated UIs and clean REST APIs. HTML/CSS/JavaScript on the front, modular Python services behind — architected for load times and cross-device consistency.
OWASP Top 10 and threat modeling by default. Smart-contract auditing (reentrancy, access control, gas misuse), parameter fuzzing for SQLi/XSS, and strict input validation to shrink the attack surface.
▌
Modular Python backends, distributed web crawlers, and data pipelines with per-stage validation and automated reporting — indexed in ElasticSearch for real-time querying.
Ethereum smart-contract development and security auditing at IBM — secure coding, iterative testing, and logic-flaw elimination before deployment.
// 05 — core
01/05 · AI threat-intelligence platform & vulnerability scanner
AI-powered cybersecurity platform with a real-time global threat dashboard. Visualises active attack paths on a 3D globe, auto-classifies severity (Low→Critical), and streams live threat feeds for DDoS, ransomware, and phishing events. Full vulnerability scanner, attack-surface mapper, phishing detector, and API security analyser — backed by CivicShield AI for scan-derived intelligence insights. Hack Energy 2.0 Finalist.
02/05 · End-to-end encrypted AI chat
Encrypted AI chat where every message is AES-256-GCM ciphered before leaving the browser — zero plaintext in transit. Session keys are derived via PBKDF2 and integrity is guaranteed by HMAC-SHA512. The Groq Llama 3.3 70B backend runs server-side with the API key fully protected. Tracks encrypted/decrypted message counts, token usage, and HMAC checks per session.
03/05 · Automated Solidity smart-contract security auditor
Static analysis engine that parses Solidity smart contracts, extracts the AST, and runs a suite of vulnerability detectors — reentrancy, integer overflow/underflow, unchecked external calls, improper access control, and gas-griefing vectors. Generates structured audit reports with per-finding severity ratings (Critical → Low) and remediation guidance. Built on the blockchain security knowledge from the IBM auditing engagement.
04/05 · ML pipeline predicting 2026 road-accident severity & blackspots
End-to-end machine learning pipeline that ingests road-accident records and predicts 2026 accident occurrences and geographic blackspots — high-risk locations where collisions are statistically concentrated. Exploratory data analysis surfaces hidden patterns; feature engineering extracts time-of-day, weather, road-condition, and vehicle-type signals. The trained classifier outputs risk-tier scores with per-feature importance breakdowns and pinpoints accident-prone zones for targeted intervention.
05/05 · Enterprise static security scanner for banking systems
Production-grade security platform built for PSB Hackathon 2026 (UCO Bank × IIT Kharagpur). Detects vulnerable dependencies, hardcoded secrets, weak cryptography, injection flaws, and insecure configurations in Python projects. Ships a full Next.js dashboard with 5-tier RBAC, immutable audit logging, and a CI/CD gate that exits non-zero on Critical findings.
// 06 — vault
Security is not a feature you bolt on — it is the substrate everything else runs on. Every system I ship assumes hostile input, least privilege, and an audit trail.
2024
Frontend Developer — 1Stop.ai
2025
Blockchain Developer — IBM
2026
Google Cybersecurity Professional — Coursera
2026
Hack Energy 2.0 Finalist — CivicShield
2027
B.Tech CSE — VIT (expected)
// live shell — type a command
divyansh@void:~/vault
// 07 — logs
Field notes from the build — multi-agent AI, WebGL performance, and offensive security. Every post traces back to something I actually shipped.
read all logs80k WebGL particles, GSAP scroll choreography, and a hydration mismatch that silently made the entire site unclickable — a debugging war story.
What actually breaks when you wire LLMs into a backend — contract-driven service boundaries, local-model fallbacks, and how to make agents testable.
How CivicShield crawls a target, extracts endpoints and parameters from JavaScript, and fuzzes them for injection flaws — offensive security, end to end.
// 08 — signal
Open to internships, security work, and ambitious builds. Signal travels fastest by email — the form transmits straight to my inbox.
divyanshg2602@gmail.com
