Building AI systems, cybersecurity tools, and research-driven software that solve real-world problems at scale.
Click a tile to dive deeper
Who I am & what drives me
14+ projects built with passion
6+ industry credentials
Published academic work
Competition achievements
Professional internships
Engineering deep-dives & research articles
Let's connect & collaborate
Automatically pulled from my repositories
Hey, I'm Jitraj Esh 👋
I'm a BCA (Cybersecurity) student, researcher, and builder based in Bengaluru, India. My work spans artificial intelligence, cybersecurity, machine learning, and large-scale software systems, with a strong focus on solving real-world problems through research and engineering.
I've built projects ranging from intrusion detection systems, LLM oversight platforms, reinforcement learning environments, and malware analysis tools to financial risk engines and social-impact applications. I enjoy turning complex ideas into practical, scalable solutions that create measurable impact.
Beyond development, I'm actively involved in research — author of a published paper presented at ICCI 2026 and currently working on a second manuscript exploring associative memory systems and precision control in AI. My interests include AI safety, machine learning research, cybersecurity, distributed systems, and intelligent agents.
When I'm not coding or researching, you'll find me participating in hackathons, contributing to open-source projects, exploring emerging technologies, and continuously learning through hands-on experimentation.
Institutional-grade risk management platform that calculates portfolio VaR and CVaR using multiple methodologies, regresses assets against macro factors, and simulates historical crises.
Machine-learning-driven decision support system that evaluates, rates, and explains hypothetical Mergers & Acquisitions (M&A) transactions using XGBoost, SHAP explainability, and historical comparables.
Real-time financial risk monitoring engine with Kafka event streaming, anomaly detection, and stress-testing for simulated market scenarios.
Enterprise-scale intrusion detection platform using ML ensembles, behavioral analytics, and MITRE ATT&CK-based threat reconstruction.
LLM oversight platform for hallucination detection, RLHF-style reward modeling, and AI safety evaluation across multi-agent systems.
Reinforcement-learning environment that optimizes cloud costs, SLA compliance, and carbon efficiency using multi-objective rewards.
Explainable malware analysis system using machine learning, anomaly detection, and SHAP-based threat attribution.
Novel associative-memory research exploring query-conditioned precision control without retraining frozen neural weights.
ML-based malware analysis tool evaluating suspicious files and generating threat risk scores with summarized security reports.
Connects volunteers with people in need during emergencies. Features emergency callback and real-time coordination.
Community web platform for uploading, browsing, and organizing images in a board-style interface.
Lightweight productivity tool for task tracking and workflow organization.
Python honeypot simulating vulnerable SSH/HTTP services to detect and log malicious login attempts for research.
System monitoring & keylogging tool capturing keyboard activity for security analysis.
Cisco Networking Academy – 2026
Foundational training in cybersecurity concepts including online safety, cyber threats, vulnerabilities, and security practices.
View Certificate →Cisco Networking Academy – 2026
Network security, threat mitigation, vulnerability assessment, risk management, and incident response fundamentals.
View Certificate →NIIT Foundation – 2026
Cybersecurity training focused on essential security practices and awareness under the Cyber Suraksha initiative.
View Certificate →IBM – 2026
Training in security operations, threat detection, incident response, SIEM concepts, and vulnerability management.
Google – 2026
Training in security foundations, threat detection, incident response, network security, Linux, SQL, and use of SIEM tools for security operations.
Presented at International Conference on Computational Intelligence (ICCI) – 2026
This research explores advancements, challenges, and societal considerations of invasive brain–computer interface technologies. It examines neural interfaces, applications in medicine and computing, ethical implications, and future directions.
Independent Research · 2026
Investigates whether precision in associative-memory retrieval can be controlled through query conditioning without retraining frozen weights. The research introduces a novel precision-control mechanism, conducts multi-seed ablation studies, and demonstrates statistically validated improvements over baseline retrieval performance while proving structural limits of the benchmark.
Selected among the Top 20 teams out of 3,200+ participants in Anvil 2026, a national-level AI and software engineering hackathon organized by Scaler. Built and presented an innovative technical solution under competitive evaluation and rapid development constraints.
View Certificate →Reached the finalist stage among 32,000+ participants in the Meta AI × Hugging Face OpenEnv Hackathon. Worked on open-source AI systems, agentic workflows, and large-scale collaborative innovation within the global AI developer community.
View Certificate →8-hour national-level hackathon at Kristu Jayanti University in collaboration with the UiPath Community, focused on AI-driven automation using agentic AI.
See Certificate →Competitive hackathon organized by Physics Wallah Institute of Innovation, solving real-world technical challenges with cross-institutional collaboration.
View Certificate →36-hour hackathon by Replit and Polaris School of Technology focusing on rapid prototyping, collaborative development, and innovative software solutions.
View Certificate →Kristu Jayanti University
Jul 2025 – Present
Reduced attendee onboarding wait times by 30% across 1,500+ records by automating verification workflows and coordinating a 10-member volunteer team. Delivered analytics dashboards across 6 event categories and streamlined certification operations for a 60-student cohort across Cisco, Oracle, MongoDB, Infosys Springboard, and NASSCOM programs.
Child Cancer Trust (NGO) · Remote
Jan 2025
Built Tableau dashboards tracking donor engagement across multiple social platforms, surfacing campaign-performance insights that supported leadership reporting, outreach strategy, and data-driven decision making.
A deep dive into MissionCtrl, a multi-agent AI oversight system designed to detect hallucination cascades and supervise fleets of collaborating AI agents. The article covers environment design, reinforcement learning with GRPO, hallucination injection strategies, reward modeling, and lessons learned while building an AI that evaluates other AIs.
An engineering-focused article documenting the design of RiskEngine and ScenarioForge, an event-driven financial risk scoring platform capable of processing hundreds of events per second using Kafka, ONNX Runtime, TimescaleDB, and Spring Boot.
An end-to-end machine learning system that quantitatively evaluates M&A transactions, fusing an XGBoost classifier trained on 70+ historical SEC EDGAR deals with expert business-logic overlays in log-odds space and full SHAP explainability. The article covers feature engineering, continuous penalty modeling, mathematically proven SHAP additivity, and a historical comparables retrieval engine.
I'm always open to new opportunities, collaborations, and conversations. Feel free to reach out!