# Avash Pati

Avash Pati is a remote worker on Himalayas, PortSwigger Web Security Academy at PortSwigger, based in India.

Source: https://himalayas.app/@avashpati

```text
Aspiring cybersecurity and AI security researcher focused on adversarially robust systems.
```

| Detail | Value |
| --- | --- |
| Availability | Open to opportunities |
| Location | India |
| Authorized to work in | India |
| Job categories | Cybersecurity Researcher, Cybersecurity Researcher, AI Security Engineer, Penetration Tester, Security Automation Engineer, Adversarial Machine Learning, Cybersecurity Research, Security Researcher, Cybersecurity Research Engineer, Threat Researcher |
| Skills | Agentic AI, Python, C#, Deep Learning, Algorithm Design, Low Level Memory Logic, PyTorch, Pandas, NumPy, Anthropic Claude API, Model Context Protocol (MCP), N8n, LangGraph, Wireshark, OverTheWire Bandit, Web Application Testing, PicoCTF, CyberChef, TryHackMe Pre Security, Xception, Out Of Distribution Testing, Adversarial Machine Learning, SQL Injection, XSS, IP Packet Analysis, Linux, SQL, Linux Administrator |

## What I'm looking for

```text
I’m looking for a role where I can build and evaluate secure AI/agent systems—privacy-first orchestration, strong observability, and rigorous defensive testing. I want to contribute to threat modeling, adversarial evaluation, and safer deployment practices.
```

## About

```text
I’m building toward a career in cybersecurity and AI security, with strong momentum from my B.Sc. Cybersecurity and B.Sc. Computer Science admissions offers. My focus spans cryptographic protocols, network security, and secure software architectures, backed by rigorous competitive mathematics and algorithmic problem solving.

I’ve pursued hands-on security work through independent research and practical training: I completed 50+ hours of advanced coursework on generative AI agent security, where I built Python client-host-server architectures using Model Context Protocol (MCP). I also studied exploit vectors in multi-agent frameworks—context leakage, prompt injection vectors, and the “confused deputy” problem—while evaluating human-in-the-loop patterns and observability telemetry for operational security.

On the defense and evaluation side, I’ve conducted adversarial machine learning research, including out-of-distribution failure analysis using Xception-based image classifiers and PyTorch-based stress testing to expose deep learning vulnerabilities. I’ve also designed privacy-first automation pipelines with n8n and LangGraph to avoid external third-party API dependencies, and I’ve validated foundational skills through TryHackMe and PortSwigger web security training.
```

## Experience

### PortSwigger Web Security Academy

PortSwigger. Jun 2026 - Jun 2026

```text
Completed apprentice-level web application security practice, including targets focused on SQL injection extraction and cross-site scripting payload execution.
```

Skills: SQL-Injection, Web-Application-Security

### TryHackMe Certified Pathways

[TryHackMe](https://himalayas.app/companies/tryhackme). Jun 2026 - Jun 2026

```text
Completed an online introduction to cybersecurity and pre-security pathway, including practical labs focused on networking and packet analysis.
```

Skills: TCP-IP, Wireshark, Defensive-Security

### B.Sc. Computer Science

[The University of Manchester](https://himalayas.app/companies/university-of-manchester). Jun 2026 - Present

```text
Secured an official undergraduate offer for a pure Computer Science program based on competitive computational logic and technical profile strength.
```

### B.Sc. Cybersecurity

University of Warwick. Jun 2026 - Present

```text
Accepted as an undergraduate offer for advanced study in cybersecurity, including cryptographic protocols, network security, and secure software system architectures.
```

Skills: Cryptographic-Protocols, Network-Security, Security-Architecture

### B.Sc. Computer Science

University of Edinburgh. Jun 2026 - Present

```text
Secured an official undergraduate offer for a pure Computer Science program based on competitive computational logic and technical profile strength.
```

### MATS Program (Pending)

Machine Learning Alignment Theory Scholars (MATS)

```text
Application is under consideration for the Machine Learning Alignment Theory Scholars (MATS) program.
```

### Algoverse Research Fellowship

Algoverse

```text
Accepted as a research fellow to engineer technical mathematical frameworks.
```

### Pivotal AI Safety Cohort

Pivotal AI

```text
Application is under consideration for a Pivotal AI safety cohort.
```

Skills: AI-Safety

### LEAF Program (AI Safety)

BlueDot Impact

```text
Enrolled in advanced AI safety tracks evaluating systemic risk, catastrophic failure mitigation, and alignment policy for frontier-scale systems.
```

Skills: AI-Safety, Systemic-Risk, Catastrophic-Failure-Mitigation, AI-Ethics

### LangGraph Secure Orchestration

LangGraph

```text
Designed self-hosted automation and agent orchestration workflows using LangGraph (and n8n), with safety policies intended to reduce data-privacy and supply-chain exploit risks.
```

Skills: LangGraph, N8n, Agent-orchestration, Offline-Security

### IBM SkillsBuild Certification

IBM SkillsBuild

```text
Earned certifications covering foundational machine learning concepts, prompt engineering robustness, and responsible bias alignment deployment.
```

Skills: Responsible-Bias

### Anthropic Academy Coursework

[Anthropic Academy](https://himalayas.app/companies/anthropic)

```text
Completed 50+ hours of advanced coursework on deploying, orchestrating, and structurally securing generative AI agent and subagent systems, including MCP-based architectures and multi-agent exploit vectors.
```

Skills: Model-Context-Protocol-(MCP), Multi-Agent-Systems, Prompt-Injection, Context-Leakage, Confused-Deputy

### picoCTF Web Penetration

picoCTF

```text
Completed web application penetration testing practice using picoCTF.
```

Skills: Web-Application-Testing

### CyberChef Cryptographic Analysis

CyberChef

```text
Performed applied cryptographic analysis using CyberChef.
```

### Google AI Essentials Certification

Google AI Essentials

```text
Completed certifications focused on foundational machine learning parameters and responsible bias alignment deployment.
```

Skills: Responsible-Bias

### OverTheWire Bandit Practice

OverTheWire

```text
Practiced Linux system administration using the OverTheWire Bandit environment.
```

Skills: Linux-Administrator

### Apart Research Alignment Jam

Apart Research

```text
Registered participant for an upcoming global AI safety code sprint (Apart Research Alignment Jam).
```

Skills: AI-Safety, Secure-Coding

## Education

### DAV Public School, Chandrasekharpur

CBSE Class 12 (Higher Secondary), Mathematics. Grade: Class 12: 86%; Grade 10: 94.6%

Activities and societies: Relevant focus: advanced competitive mathematics & algorithmic problem solving (calculus, probability, linear algebra, mathematical logic).

```text
Completed CBSE Class 12 with a score of 86% and previously achieved a 94.6% score in Grade 10. Focused on advanced competitive mathematics and algorithmic problem solving.
```

### University of Manchester

Bachelor of Science in Computer Science, Computer Science. 2026 - Present

```text
Received an official undergraduate offer for a B.Sc. in Computer Science starting June 2026. Admission is based on competitive computational logic and the applicant’s technical profile.
```

### University of Edinburgh

Bachelor of Science in Computer Science, Computer Science. 2026 - Present

```text
Received an official undergraduate offer for a B.Sc. in Computer Science starting June 2026. Admission is based on competitive computational logic and the applicant’s technical profile.
```

### University of Warwick

Bachelor of Science in Cybersecurity, Cybersecurity. 2026 - Present

```text
Received an official undergraduate offer for a B.Sc. in Cybersecurity starting June 2026. Coverage is focused on cryptographic protocols, network security, and secure software system architectures.
```

## Tech stack

- NumPy
- Pandas
- Gmail
- PyTorch
- Wireshark
- Linux
- SQL
- Anthropic Claude API
- Score
- N8N
- Agentic
- LangGraph
- Model Context Protocol (MCP)

To contact Avash, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [adarsh vh](https://himalayas.app/@adarshvh): Machine Learning Intern, Cognifyz Technologies
- [Smit Borkhetaria](https://himalayas.app/@smitborkhetaria): Cyber Security Researcher, Advanced Cyber Research Lab
- [Ujjwalpreet Singh](https://himalayas.app/@ujjwalpreetsingh): Cybersecurity Team Lead, Google Developer Group (GDG GTBIT)
- [D RUDRA PRATAP](https://himalayas.app/@drudrapratap): Cyber Security Engineer, Edysor.ai
- [Parth Bhatt](https://himalayas.app/@parthbhatt3): Cyber Security Intern, DRDO (SAG), Ministry of Defence
- [yashash mathur](https://himalayas.app/@yashashmathur): Club Head, AI & Cybersecurity, AEGIS, RVITM
- [Reeti Agarwal](https://himalayas.app/@reetiagarwal): Vice President Cyber Blockchain Club, Symbiosis Institute of Technology
- [Sanchay Jain](https://himalayas.app/@sanchayjain): Cybersecurity Intern, Uttar Pradesh State University of Forensic Science
- [Mir Arsalan](https://himalayas.app/@mirarsalan): Independent Cybersecurity Researcher, Self Directed
- [Ishan Shishodiya](https://himalayas.app/@ishanshishodiya): Teaching Assistant – Deep Learning & Generative AI, IIT Madras BS Degree Programme
