Harshdeep Athawale profile

Harshdeep Athawale

Security Researcher/Engineer

I build software and break web and mobile apps to report critical bugs - 35+ vulnerability reports across Red Bull, Goldman Sachs, Adobe, Netflix, NVIDIA, Anduril, Flipkart, and Coca-Cola.

Open to security internships · Summer 2027

Disclosure pipeline
Recon

API & GraphQL

FindingsLive

35+ vulnerability reports

9.1CVSS

Critical GraphQL flaw exposing employee PII at Red Bull

Tech Stack[01]

Experience[02]

  1. HackerOne logo

    HackerOne

    Current

    Security Researcher · Part-time

    Dec 25 - Present

    San Francisco, CA (Remote)

  2. Google Developer Groups TIET logo

    Google Developer Groups TIET

    Current

    Full-time

    Jul 25 - Present

    Patiala, IN (On-site)

    1. Head of Cyber Security

      Aug 26 - Present

    2. Core Member

      Jul 25 - Aug 26

  3. Iris Intelligence logo

    Iris Intelligence

    Security Engineer · Internship

    Apr 26 - Jun 26

    Delhi, IN (Hybrid)

Honors & Awards[03]

View all

Cyber AI Hackathon 2026

Innovation Award

2026

Oct

Internal Smart India Hackathon (SIH) 2026

Judge

2026

Sep 12

NioHack 2026

Winner · 1st Place

2026

Sep 5-6

PSB Hackathon Series 2026

National 1st Runner-Up

2026

Aug 1

Featured Projects[04]

View all
WAF · inspecting96% detection

A reverse-proxy WAF combining sliding-window rate limiting with transformer-based request inspection, benchmarked against ModSecurity with OWASP CRS v4 on 675 labeled real-world payloads for a 96% detection rate vs. 46%. Fine-tuned DistilBERT to detect 10 web attack classes - Path Traversal, SQLi, XSS, SSRF, and IDOR - at 97.5% accuracy, with OWASP Juice Shop, WebGoat, and DVWA deployed behind a Docker Compose reverse-proxy gateway enforcing per-IP rate limiting and monitor, block, and challenge modes before reaching the origin.

  • PyTorch
  • DistilBERT
  • FastAPI
  • Next.js
  • Docker
  • Redis
  • Nginx
Deepfake · 4-agent scanScanning

Full-stack AI platform with a 4-agent orchestration system (Perception, Detection, Compression, Cognitive) for deepfake media analysis. ResNet50/EfficientNet models trained on FaceForensics++ (16GB, 7000+ samples) for high-accuracy detection of AI-generated content.

  • Next.js
  • React
  • TypeScript
  • Node.js
  • MongoDB
  • FFmpeg

GitHub Activity[05]

@HarshdeepAthawale
Contributions
2,721past year
Longest streak
21days
Current streak
8days
Best day
1001on Oct 5
Less01-23-56-1011+More
“

Complexity is the worst enemy of security.

- Bruce Schneier

Counting visitors...