Experience

Research Experience

Scientist (Research and Technology) at the Luxembourg Institute of Science and Technology, Luxembourg since March 2024 (at present)

Aayush continues to develop new methods employing Artificial Intelligence (AI) to improve software security by focusing on software vulnerabilities, both real (existing) and synthetically generated.
He is exploring the capabilities of Large Language Models (LLMs) in implementing software self-healing, focusing on automatically generating source code patching vulnerabilities.
He is developing automated solutions to identify the attack vectors (in the context of a 5G-core network) to prevent unauthorized entrance and extraction of telecom data.
Aayush employed Deep Learning (DL), primarily Transformers architecture, to automatically identify API fuzzing attack patterns, in order to tackle cyber attacks utilizing Fuzzing.
He is developing automated assurance techniques to ensure the quality of Generative AI-enabled educational activities.
Aayush is designing a semi-automated framework to transform real-world threat intelligence into executable cyber range scenarios, with the goal of delivering realistic, threat-informed, and role-specific cybersecurity training exercises.

Aayush designed solutions harnessing AI, especially Deep Learning to automatically define and learn features from source code representation replacing 100% manual effort.
He trained Neural Networks and Machine Learning models to achieve task-specific classifications with up to 87% Precision and Recall.
He employed LLMs and GPTs to induce artificial faults enabling up to 65% increased test effectiveness in detecting zero-day vulnerabilities.

Aayush conducted lectures on the “Introduction to Machine and Deep Learning” to Bachelors in Computer Science (BICS)
Lecture series (videos) on the “Introduction to Machine and Deep Learning” are available here. The videos were uploaded to make lectures available to students during the COVID-19 lockdown period.
He also taught “Introduction to Software Testing” to Bachelors in Applied Information Technology (BINFO)
Lecture series (videos) on the “Introduction to Software Testing” are available here. The videos were uploaded to make lectures available to students during the COVID-19 lockdown period.

Conference Track Chair & Program Committee Member

Aayush served/is serving as the Chair/Program Committee Member/Reviewer for the following Conference tracks:

Peer Reviewer (Journals)

Aayush continues to serve as a reviewer for the following Academic Journals:

Senior Software Engineer at Futures First, India from September 2014 to February 2017

Aayush developed multi-threaded Windows applications to process and display prices and trade updates of invested stocks with a 3-second refresh rate.
He implemented microservices to capture up to 12 stock price updates per second (real-time) via market APIs.
He engineered a Futures and Commodities trading platform empowering min. 95% organizational market investments.

Senior Software Engineer at Indus Valley Partners, USA from August 2012 to September 2014

Aayush built investment-compliance applications to automate legal due diligence auditing replacing up to 85% manual effort.
He implemented microservices to enable easily configurable email alerts streamlining min. 80% business process workflow.
He integrated portfolio dashboards and reporting capabilities for 11 Business Experts (SMEs) to facilitate overall debt investment overview.

Associate Software Engineer at Fiserv, USA from May 2010 to August 2012

Aayush crafted web applications to streamline investment portfolio metrics capturing, resulting in a 53% increased efficiency.
He resolved defects and maintained source code to promote the management of 3 large-scale Electronic Fund Transfer banking projects.
He ensured high quality (min. 85% defect-free) of applications under development through peer-reviewing code modifications.

Causal Analysis and Resolution coordinator at Fiserv, USA from May 2010 to August 2012

Aayush performed Defect Root Cause analysis and processed inefficiencies for 95% of Business Unit projects.
He executed corrective actions to drive continuous process enhancements, to maintain min. 85% defect-free software quality and to gain 15% increased efficiency in delivering modules every quarter.