CV

Education, experience, projects, skills and certificates.

Contact Information

Name Krishna Gunjan
Professional Title Computer Science undergraduate
Location Bengaluru, Karnataka

Professional Summary

Computer Science undergraduate securing Linux systems and autonomous vehicles: automated CIS hardening, guarded multi-agent LLM remediation, and spoof and anomaly detection for vehicle sensors. Preparing two projects as research papers.

Experience

  • 2026 - Present

    Bengaluru, India

    Cybersecurity, Network & Communications team
    Centre of Excellence in Autonomous Mobility, MIT Bengaluru
    • Hardened Ubuntu 22.04 LTS to CIS Benchmark v3.0.0 (Level 1 + 2 Server) with a custom Ansible playbook, raising OpenSCAP compliance from 51.5% to 100% (125 failing rules fixed).
    • Designed the campus-wide communication network linking the central server, roadside units and the autonomous vehicle, and planned its RF bands and transmit power (IN865) under WPC licence-exempt rules.
    • Led a team of over 20 freshmen through a 2-month onboarding task phase.

Education

  • Bengaluru, India

    Bachelor of Engineering
    Manipal Institute of Technology, Bengaluru (MAHE)
    Computer Science

Projects

  • 2026 - 2026
    SentinelLoop: Agentic CIS Hardening with a Human in the Loop (AgentOps'26 Hackathon)

    Python, multi-agent LLMs (Ollama, Qwen2.5-Coder), OpenSCAP, Streamlit, VirtualBox

    • Built a multi-agent LLM loop that audits Ubuntu 24.04 servers against CIS Level 1 (OpenSCAP), applies fixes, verifies each one, and rolls back any that break SSH, sudo or required services.
    • Restricted the AI to 15 allowlisted actions behind a scope lock; guardrails blocked 79 unsafe proposals.
    • Added a Streamlit approval dashboard with live dry-run diffs, a persistent undo journal and pre-patch VM snapshots.
    • On a 3-VM lab, raised CIS compliance from 66.3% to 92% on every host and cut open ports from 6 to 2 or fewer.
    • Cut known vulnerabilities from 4 to 0 on the web server and 4 to 2 on the database server; a Dirty Pipe (CVE-2022-0847) exploit that worked before patching failed after.
  • Anomaly Detection for Autonomous Vehicles

    Python, CARLA, TensorFlow/Keras, scikit-learn, Redis

    • Trained a 22-feature autoencoder (22-16-8-4-8-16-22) on normal CARLA telemetry (position, velocity, IMU, GNSS, controls) to flag anomalies and traffic-law violations live at a 95th-percentile threshold.
    • Ran a live demo in CARLA showcasing ADAS system behavior and applications in detecting traffic law violations by running through a red light, bypassing speed limit, etc., with 97.9% accuracy.
  • Other projects

    Cybersecurity home lab (5 VMs, SIEM monitoring), portfolio-rag (local-LLM RAG portfolio adviser), Cancer Analysis (Streamlit ML app), BeatPrints (open-source contributor), Shell and Sneak (Python)

Skills

Security (): CIS Benchmarks, OpenSCAP, Ansible, Wireshark, Splunk (SIEM), Microsoft Entra ID, Tailscale/WireGuard
AI/ML (): PyTorch, TensorFlow/Keras, scikit-learn, NumPy, pandas, local LLMs and agents (Ollama), RAG
Vehicles & Networks (): CAN bus (python-can, SocketCAN), mmWave radar, RF planning, CARLA
Languages & Platforms (): Python, C, Java, Bash, Go, SQL; Linux, VMware, VirtualBox, Azure, Redis, Git

Certificates

  • OCI 2025 AI Foundations Associate - Oracle
  • Foundations of Cybersecurity - Google
  • Deloitte Data Analytics Job Simulation - Forage
  • AI Skills Fest 2026 - Microsoft
  • Software Engineer Intern - Hackerrank
  • Networking fundamentals - Packt