Harsh M. profile photo 2 on Tal
Harsh M. profile photo 1 on Tal
Harsh M. profile photo 2 on Tal
Harsh M. profile photo 1 on Tal

Harsh M.

Building AI security tooling that finds what manual scans miss | Cybersecurity @ Tech Mahindra | RAG, LLM Systems, Backend Security | IIT Mandi × GEC Rajkot ’26

India

< 1 year of work experience

Tech Mahindra
Aug 20261 moPresent

Bengaluru

- Automated a manual cloud security-reporting workflow via a custom command-line script — auto-tracking report folders and eliminating manual renaming, cutting task time by ~50%. - Fetched and consolidated multi-domain cloud security reports (vulnerabilities, secrets, malware, network exposure) from a Ubuntu VM using CNAPP tooling. - Cleared CA TDM Level 2 certification in software testing methodologies.
CODTECH IT SOLUTION
Jan 20264 moMay 2026

Telangana, India

Harsh’s AI Stack

claude code
claude code
codex
codex
deepseek
deepseek
gemini
gemini
lovable
lovable
manus
manus

Education

Indian Institute of Technology, Mandi

Cyber Security and Ethical Hacking

Indian Institute of Technology, Mandi

20251 yr2026
Government Engineering College, Rajkot

Bachelor of Engineering - BE, Computer Engineering

Government Engineering College, Rajkot

20224 yrs2026

Join Harsh and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Harsh M. on Tal

Headline

Building AI security tooling that finds what manual scans miss | Cybersecurity @ Tech Mahindra | RAG, LLM Systems, Backend Security | IIT Mandi × GEC Rajkot ’26

About

I build the systems most "AI security" tools skip — the RAG pipeline underneath, the scanner that actually catches SQLi and SSRF, the backend that survives real load. Computer Engineering grad, GEC Rajkot. Minor in Cybersecurity & Ethical Hacking, IIT Mandi. Currently in Enterprise AI & Cybersecurity at Tech Mahindra, Bangalore. What I've shipped: — DocuMind: a full RAG pipeline (chunking, vector embeddings, ChromaDB) with dual-provider LLM support (Gemini + OpenAI) — CyberGuard Pro: a multi-threaded scanner (200 concurrent ports), SSRF prevention, and automated PDF audit reporting — OrderDash: a Spring Boot + PostgreSQL backend with JWT auth and 40+ endpoints — Sentinel: a Redis-backed rate limiter using the token bucket algorithm I think the next interesting AI infrastructure won't come out of well-resourced labs — it'll come from people who had to build it with fewer resources and tighter constraints. That's not a complaint. That's just what constraints do to how you think. Specialties: RAG pipelines, LLM integration, vulnerability scanning, backend system design, Java, Python, Spring Boot, FastAPI. Building at the intersection of AI and security, or want to talk distributed architecture and LLMs at scale? DM me.

Location

India

AI stack

claude code

codex

deepseek

gemini

lovable

manus

Work experience

Cyber Security Intern at Tech Mahindra, Aug 2026 - Present · 1 mo, Bengaluru, - Automated a manual cloud security-reporting workflow via a custom command-line script — auto-tracking report folders and eliminating manual renaming, cutting task time by ~50%. - Fetched and consolidated multi-domain cloud security reports (vulnerabilities, secrets, malware, network exposure) from a Ubuntu VM using CNAPP tooling. - Cleared CA TDM Level 2 certification in software testing methodologies.

Cyber Security Trainer at CODTECH IT SOLUTION, Jan 2026 - May 2026 · 4 mo, Telangana, India

Education

Cyber Security and Ethical Hacking, Indian Institute of Technology, Mandi

Bachelor of Engineering - BE, Computer Engineering, Government Engineering College, Rajkot

Public links

github: https://github.com/hrshmakwana

leetcode: https://leetcode.com/hrshmakwana

Harsh M. - Cyber Security Intern in India | Tal