Headline

BTech CSE student at PESU | Gen AI Engineering Intern @ Kahuna Labs| Ex Cybersecurity and ML R&D Intern @ CMC, IIIT Hyderabad | Passionate About Cybersecurity and Machine Learning

About

A Computer Science undergraduate at PES University, expected to graduate in 2027, with hands-on experience in cybersecurity and machine learning. Currently an intern at Cyber Manthan Center, IIIT Hyderabad, working on conceptualizing and developing a framework for Security Operation Centre (SOC) automation and co-authoring a paper on SOC methodologies. Past roles include key contributions at SecPod and C3iHub, IIT Kanpur, focusing on cybersecurity research and development. As part of the university’s CTF team, supported the organization of a national-level competition and contributed to challenge design and team mentorship. Skilled in cybersecurity, deep learning, and vulnerability assessment.

Location

Bengaluru

AI stack

cursor

gemini

hugging face

langchain

claudecode

Qdrant

Work experience

Gen AI Engineering Intern at Kahuna Labs, Jun 2026 - Present · 2 mo, Bengaluru, Karnataka, India, • Building a golden test dataset for an AI-powered support automation platform, curating production support cases to serve as ground-truth benchmarks for evaluation. • Engineering an evaluation framework infrastructure to measure and compare LLM output quality against the golden dataset, enabling data driven iteration on prompt design. • Implementing end-to-end validation pipelines for the multi-agent LLM system to ensure reliability across heterogeneous customer support workflows.

Cybersecurity Research and Machine Learning Intern at Cyber Manthan Center, IIIT Hyderabad, Dec 2025 - Mar 2026 · 3 mo, Hyderabad, Telangana, India, • Architecting an Agentic SOC framework to alleviate analyst burnout by orchestrating autonomous LLM-agents via LangGraph, enabling the system to “pre-digest” high-volume telemetry before human review. • Designed a multi-agent pipeline that ingests heterogeneous alerts (Wazuh, Suricata, Zeek), normalises them into a unified schema, and routes them to three specialised agents for anomaly scoring, threat correlation, and natural-language session summarisation. • Engineered a discriminator-based feedback loop with structured output validation to enforce schema consistency across LLM outputs, with downstream scoring and remediation agents producing prioritised summaries for analysts.

Cybersecurity Research and Development Intern at SecPod, Aug 2025 - Dec 2025 · 4 mo, Bengaluru, Karnataka, India, Expanded threat intelligence coverage by engineering a custom data ingestion schema using Apache Airflow that normalizes heterogeneous threat feeds, directly improving the downstream detection capabilities of the company’s vulnerability management product.

Cybersecurity Research and Development Intern at C3iHub, IIT Kanpur, Jun 2025 - Sep 2025 · 3 mo, Kanpur, Uttar Pradesh, India, • Eliminated around 60% of IDS alert volume by engineering a stateless pre-filtering mechanism that isolates and discards mechanically stable benign noise before it reaches the SIEM. • Benchmarked the filter against expert analyst triage, achieving 99% recall and validating that ML-driven pre-filtering matches human accuracy at scale.

Education

PES University

Machine Learning Summer School, University of Oxford

Public links

github: https://github.com/kshitijkota