Ankur Lahiry

Ankur Lahiry

ML Systems, Graph Learning, and Explainable AI Ph.D. Student, Texas State University

Hello! I’m Ankur Lahiry, a Ph.D. student in the Department of Computer Science at Texas State University, working in the Per4ML Lab under the supervision of Dr. Tanzima Islam.

I apply machine learning to improve performance, efficiency, and scalability in complex computing environments. My work centers on graph learning, decision intelligence, causal analysis, and explainable AI, with HPC and GPU systems serving as the applied domain for building and validating practical ML systems.

📝 Career Summary

  • 🎓 Ph.D. Student in Computer Science focused on machine learning systems, graph-based modeling, and explainable AI at Texas State University.
  • 📈 Expertise in performance modeling, representation learning, causal reasoning, and scalable ML pipelines for large operational datasets.
  • 🤝 Collaborator with Brookhaven and Argonne National Laboratories, applying ML to performance analytics, reliability analysis, and anomaly diagnosis in HPC environments.
  • 🧠 Focused on building interpretable, production-relevant ML systems for complex, high-volume computing data.
  • 💻 5+ years of professional experience in software engineering, with a focus on building large-scale social communication platforms and developing secure, real-time digital payment solutions.

🧠 Research Interests

My research interests lie at the intersection of machine learning, graph-based reasoning, and large-scale systems data.

  • Machine Learning (ML):
    • Designing adaptive, explainable, and data-driven models.
    • Building learning systems that remain reliable on large, noisy, operational datasets.
  • Graph Representation Learning:
    • Building graph-based models to capture complex relationships in system performance data.
    • Applying GNNs for anomaly classification, behavior modeling, and structured prediction.
  • Decision Making & Explainable AI:
    • Developing decision-support systems for navigating performance trade-offs in complex environments.
    • Making ML-driven recommendations transparent and interpretable.
  • Causal Modeling & Anomaly Detection:
    • Designing methods for early detection, root-cause discovery, and robust diagnosis.
    • Moving beyond correlation toward more reliable explanations of system behavior.
  • Applied ML for HPC and GPU Systems:
    • Using HPC and GPU traces as real-world testbeds for scalable ML, graph learning, and decision intelligence.
    • Developing practical monitoring and analytics workflows for large computing environments.
  • Synthetic Performance Data Generation:
    • Generating realistic synthetic traces to augment training data for performance models.
  • Human–Computer Interaction (HCI):
    • Creating intuitive interfaces to bridge human insight with ML-powered systems.

🌟 My Vision

I’m passionate about bridging machine learning research and real-world applications.
My vision is to build scalable, interpretable, and decision-aware ML systems that work reliably in complex operational settings and make advanced analytics more transparent and useful.

I aim to contribute to high-impact work in machine learning systems, explainable AI, graph learning, and intelligent analytics, with HPC and GPU platforms serving as strong application domains for validating these ideas at scale.

📚 Academic & Professional Background

  • Aug 2022 – Present: Ph.D. in Computer Science, Texas State University (Expected Graduation: Summer 2027)

  • Spring 2023 – Present: Doctoral Research Assistant, Per4ML Lab, Texas State University
    • Developed a unified decision-intelligence system for HPC environments to recommend performant configurations by balancing speed, cost, and reliability trade-offs with explainable outputs and uncertainty-aware ranking, scaling to traces with 1.3B samples (126 GB) and achieving up to 100× faster training and 80× faster inference than state-of-the-art generative baselines.
    • Design graph-based and explainable AI methods that turn complex system logs into intuitive signals, enabling early detection and diagnosis of unusual performance behavior in large-scale HPC systems.
    • Engineered scalable GPU log-analysis pipelines for large trace datasets using distributed partitioning and parallel processing, achieving a 67% improvement in scalability while enabling fast identification of performance variability, memory stalls, and system bottlenecks.
    • Collaborate with national laboratory partners to validate methods on large-scale clusters and integrate results into production monitoring workflows.
  • June 2023 – August 2023: Summer Research Intern, Brookhaven National Laboratory, Upton, New York
    • Worked on the Chimbuko Project, a performance analytics framework for monitoring and improving efficiency of large-scale supercomputing applications.
    • Developed a novel representation learning approach to automatically detect performance anomalies in complex computing workflows.
    • Analyzed real HPC performance data to identify inefficiencies and unusual behaviors.
  • Fall 2022: Doctoral Instructional Assistant, Texas State University
    • Assisted in teaching Computer Architecture (CS4310), supporting both lectures and lab sessions.
    • Guided and mentored undergraduate students in assignments, lab work, and term projects.
    • Collaborated with faculty to design and evaluate course assignments, ensuring alignment with learning objectives.
    • Encouraged active learning by fostering technical discussions and peer-to-peer problem solving.
  • May 2012 – Feb 2017: B.Sc. in Computer Science and Engineering, Bangladesh University of Engineering and Technology (BUET)

💼 Industry Experience Highlights

I bring over five years of hands-on software engineering experience, working with both startups and established companies in Bangladesh and the United States. My background includes mobile application development, secure financial platforms, team leadership, and product architecture.

🇧🇩 DataBird (DeshiPay & Ridmik Labs) — Software Engineer (2020–2022)

  • Spearheaded the iOS development for DeshiPay, a fintech platform enabling secure mobile transactions in Bangladesh.
  • Coordinated and scaled the iOS team to align with product goals and security requirements.
  • Implemented real-time messaging and voice-over-IP services using XMPPFramework and WebRTC for RidmikChat.
  • Worked cross-functionally with security, backend, and design teams to ensure seamless platform integration.
  • Practiced Agile methodologies including sprint planning, code reviews, and feature iteration.

🍽️ Prefeex Ltd. — Senior Software Engineer (2019–2020)

  • Led the complete development of a restaurant reservation iOS application, from design to deployment.
  • Oversaw system architecture, implemented major feature modules, and enhanced user experience.
  • Improved development workflows through Agile processes, boosting team velocity and product stability.
  • Ensured the app’s scalability to support an expanding user base and feature set.

💳 iPay Systems Ltd. — Software Engineer (2017–2019)

  • Developed the iPay iOS app, enabling secure real-time financial transactions.
  • Implemented robust security features and built iPay SDK for third-party payment integration.
  • Created internal frameworks to streamline app development and improve system maintainability.
  • Contributed to code reviews, feature enhancements, and production rollouts in a fast-paced fintech environment.

🏢 BellBizzer Inc. (Seattle, WA) — Senior Software Engineer (Remote, part time)

  • Led iOS development for a rental marketplace platform, contributing to architecture, design, and front-end modules.
  • Actively collaborated with UI/UX designers and stakeholders to build a smooth and intuitive experience.
  • Monitored app performance, optimized user interactions, and improved system reliability.
  • Played a key role in client communication and product strategy discussions to scale the platform effectively.

🧰 Technical Skills

  • Programming: C, C++, Python, Swift, Java, Go

  • Machine Learning & AI (Current Focus):
    • Model Architectures: Graph Neural Networks (GNN), Attention Mechanisms, Deep Neural Networks (DNN), Large Language Models (LLMs)
    • Core Techniques: Representation Learning, Anomaly Detection, Explainable AI (XAI), Feature Engineering, Transfer Learning
    • Applications: Performance Analytics in HPC, System Behavior Modeling, Anomaly Classification, Intelligent Monitoring Systems
  • Past Industry Expertise:
    • Databases: SQL, MySQL
    • Frameworks & Tools: PyTorch, XMPPFramework, WebRTC, Git, Agile/Scrum
    • Mobile App Development (iOS) and Cross-Platform Systems
    • Scalable Architecture Design and Product Development
    • Real-time Communication Platforms & Secure Payment Systems