Akhil V Nair

Akhil V Nair

Building intelligent systems at the intersection of AI and security

View Projects Get in Touch View Resume
5+ Projects
5+ Certifications
1 Hackathon Win

About Me

Get to know me better

AI/ML Engineer & Security AI Researcher

I am an AI/ML Engineer with a unique interdisciplinary background combining Computer Science with Criminology. This distinctive combination allows me to approach AI challenges from both technical and investigative perspectives, building intelligent systems that address real-world security threats.

As one of the winners in HAC'KP 2025, I built deep learning models for the Kerala Police CyberDome's initiative in creating a secure digital future, enhancing law enforcement capabilities with computer vision models optimized for real-time operational environments. My work focuses on security-oriented AI applications including LLM security, threat detection systems, forensics tools, and privacy-preserving machine learning.

My technical expertise spans deep learning, natural language processing, generative AI, and full-stack development, enabling me to create end-to-end ML pipelines that are production-ready and secure by design.

AI/ML Engineering

Deep learning, computer vision, NLP, and model deployment for production systems

Security AI

LLM security, threat detection, forensic analysis, and adversarial robustness

Full-Stack ML

End-to-end pipelines with FastAPI, React dashboards, and cloud deployment

Experience

My professional journey

Jan 2026 – Feb 2026

Project Trainee — Network & Cybersecurity Division

Bharat Electronics Limited (BEL), Bangalore

  • Developed AI/ML models for intrusion and anomaly detection to identify network threats and security anomalies in real-time environments.
  • Developed CLYR Hawk, a cloud-enabled multi-model ML framework for real-time network threat and zero-day attack detection with serverless inference on AWS Lambda, and a React monitoring dashboard.
Python AWS Lambda React Machine Learning Network Security
October 2025

HAC'KP 2025 Winner — AI Team Member

Kerala Police Cyberdome, Kerala

  • Enhanced the Grapnel dark web crawler into a real-time law enforcement solution for illegal content detection and takedown.
  • Built and deployed deep learning models for age estimation, NSFW classification, image captioning, and pose estimation, optimized for real-time operational performance.
Python TensorFlow Computer Vision Deep Learning Real-time Systems
Jun 2025 – Jul 2025

AI & Machine Learning Intern

Edunet Foundation (AICTE & IBM SkillsBuild), Remote

  • Built an Employee Salary Prediction model using Python, Scikit-learn, and TensorFlow, implementing an end-to-end ML pipeline from data preprocessing and feature engineering through deployment.
  • Gained hands-on experience in supervised/unsupervised learning, feature engineering, and model optimization techniques.
Python Scikit-learn TensorFlow Feature Engineering Model Deployment

Projects

Some of my recent work

ThinkSym

Neuro-Symbolic LLM Security Platform

Built a production-oriented LLM security platform addressing prompt injection, jailbreaks, PII extraction, harmful content, social engineering, and other LLM threat vectors through a 5-layer pipeline combining SmoothLLM, fine-tuned ModernBERT and Llama Guard classifiers, Scallop probabilistic reasoning, Z3 formal verification, and Neo4j graph corroboration, with an explainable FastAPI + Next.js dashboard.

Python ModernBERT Llama Guard Scallop Z3 Neo4j FastAPI Next.js
View Source

AI Image Forensics

Investigative Tool Suite

Developed a modular 8-tool forensic suite featuring perceptual image matching with pHash, SSIM, and ORB; a fine-tuned ResNet18 classifier; OCR, metadata extraction, and AI-generation heuristics; CLIP semantic search with caching and clustering; YOLO segmentation with face anonymization; and a FastAPI + React/Konva interface for manual and AI-assisted image redaction.

Python ResNet CLIP YOLOv8 FastAPI React Computer Vision
View Source

CyPrompt

AI-Driven Security Platform

Built a full-stack virtual SOC analyst using XGBoost and Autoencoder models for hybrid threat detection across 10+ attack types, with a 5-stage Gemini LLM pipeline for narrative threat analysis, SHAP-based explainability, D3.js visualizations, real-time WebSocket streaming, and role-based access.

Python XGBoost TensorFlow Google Gemini FastAPI WebSocket SHAP D3.js
View Source

Privacy-Focused Age Estimation

Client-Side Real-Time Inference

Built a fully client-side age estimation system using faceapi.js and TensorFlow.js, enabling on-device inference with no image data leaving the browser for low-latency, privacy-preserving real-time predictions.

TensorFlow.js faceapi.js JavaScript Computer Vision Privacy-First
View Source

Skills & Expertise

Technologies I work with

Programming

Python C++ JavaScript SQL Bash PowerShell

AI/ML

PyTorch TensorFlow Scikit-learn OpenCV Hugging Face

GenAI & LLMs

RAG LLM Agents Local LLM Inference

Data & Analytics

Pandas NumPy Matplotlib SHAP

Backend & APIs

FastAPI REST APIs Node.js

Cloud & Tools

AWS Docker Git PostgreSQL MongoDB

Certifications

Industry-recognized credentials

Introduction to Cybersecurity

Cisco Networking Academy

2026

Artificial Intelligence Fundamentals

IBM SkillsBuild

2025

Cybersecurity Fundamentals

IBM SkillsBuild

2025

Data Analytics

NASSCOM

2025

Cloud Computing

NPTEL – IIT Kharagpur

2025

Get In Touch

Let's work together

Email

akhilvvnair@gmail.com

Phone

+91 9526532426

Location

Thiruvananthapuram, Kerala, India

Availability

Available for AI/ML roles, research collaborations, and freelance projects.

A

Akhil's AI Assistant

Online

Hey! 👋 I'm Akhil's AI assistant. Ask me anything about his projects, skills, or experience!