About
I am driven by hard problems. The ones nobody has cracked yet are the ones I want to work on, and I have built my career around finding them.
I am a research scientist with a deep software engineering background. That combination lets me take an idea from whiteboard to prototype to production infrastructure that other teams, and now autonomous vessels, depend on. At NSWC Crane I lead and mentor a team of 8+ software engineers and research scientists and serve as the computer vision subject matter expert for Navy program offices, Technical Warrant Holders, and SBIR efforts.
I am completing a Ph.D. in Computer Science at Indiana University, researching deep neural networks for denoising and verifiable, non-black-box models, with published work spanning computer vision, representation learning, scalable clustering, and statistical relational AI.
Experience
Research Scientist & Computer Vision SME, Technical Lead
- Lead and mentor a team of 8+ software engineers and research scientists building ML Ops and evaluation infrastructure for EO/IR computer vision.
- Architected the full ML Ops pipeline for EO/IR algorithm evaluation: sensor processing, storage and compute, algorithm SDK, test harness and metadata standards, and a searchable catalog of video, metadata, and historical algorithm performance.
- Fielded the pipeline on autonomous surface vessels for real-time EO/IR perception, integrated with the Unmanned Maritime Autonomy Architecture (UMAA), and into modeling and simulation software for synthetic testing; now integrating into the Navy's Rough Casper electronic-warfare framework for EO/IR awareness.
- Drove cross-command adoption across NAVSEA, NAVAIR, and NAVWAR; contributed to a Navy-wide test harness interoperability working group on T&E standards for AI and autonomous systems.
- Led automated video target labeling research, reaching 95%+ fully autonomous labeling.
- Researched how image and video compression affects object detection and built hardware-accelerated, high-fidelity compression; custom lossless encoders cut a raw sensor archive from 2 PB to under 800 TB.
- Demonstrated a TRL-4 prototype converting natural-language prompts to segmentation masks with real-time gimbal tracking.
- Lead and draft research portfolios, roadmaps, and technology surveys for program offices, including the Portfolio Acquisition Executive for Robotic and Autonomous Systems, then execute the research; secured competitive funding for a 3-year computer vision roadmap.
- Monitor SBIR execution; author standards and white papers on perception metadata and EO/IR evaluation.
- Mentor students and collegiate teams; advised the University of Notre Dame team that won Overall Grand Champion at the 2025 AI Maritime Maneuver Indiana Collegiate Challenge.
Ph.D. Candidate & Research Assistant
Doctoral research on deep neural networks for denoising single-cell RNA-seq data and on building verifiable models rather than black-box models. Thesis proposal completed; defense scheduled November 2026.
Adjunct Professor, Computer Science
Developed and taught the undergraduate Introduction to Python course. Mentored 5+ students per semester on academic pathways, research opportunities, and careers.
Adjunct Professor, Introductory Programming
Taught introductory programming courses, building students' foundational coding and problem-solving skills.
Education
Skills
- Computer vision
- Object detectionObject trackingSegmentationVision-language modelsEO/IR sensorsVideo compressionSynthetic imagery
- ML infrastructure
- ML OpsEvaluation harnessesAutomated labelingMetadata standardsEdge deploymentTensorRT / ONNXCI/CD
- Autonomy
- Unmanned surface vesselsUMAAModeling & simulationAutonomy T&ECounter-UAS evaluation
- Languages & tools
- PythonC++CJavaCUDAPyTorchTensorFlowOpenCVDockerKubernetesLinux
- Leadership
- Technical leadershipMentoringS&T roadmapsProposal writingSBIR oversight
Selected Publications
- P. Sharma, M. Malec, et al. "Geometric-k-means: A Bound-Free Approach to Fast and Eco-Friendly k-means." Machine Learning (Springer), 2026.
- M. Malec. "An Interpretable Latent-Supervised Autoencoder for Single-Cell Representation Learning." Bioinformatics, 2026.
- M. Malec et al. "Blurred Lines: Training Object Detection Algorithms with Degraded Synthetic Datasets." MSS Parallel Proceedings, 2026.
- M. Malec et al. "Evaluating Effect of Image/Video Compression on Imagery Utility." MSS Parallel Proceedings, 2025.
- S. Koutsares et al. "Modeling and Visualization for Emission Signatures (MoVES)." MSS Parallel Proceedings, 2025.
- S. Koutsares et al. "Synthetic and Hybrid Imagery Products for Standardization (SHIPS)." MSS Parallel Proceedings, 2024.
- M. Malec, H. Kurban, M. Dalkilic. "ccImpute: An Accurate and Scalable Consensus Clustering Based Algorithm to Impute Dropout Events in Single-Cell RNA-seq Data." BMC Bioinformatics, 2022.
- M. Malec et al. "Inductive Logic Programming Meets Relational Databases: Efficient Learning of Markov Logic Networks." ILP, 2016.
Honors & Awards
- Best Student Paper Award, International Conference on Inductive Logic Programming (ILP), 2017
- Mentor, University of Notre Dame team: Overall Grand Champion, AI Maritime Maneuver Indiana Collegiate Challenge, 2025
- NSWC Crane: 2× Time Off Award, 2× On the Spot Cash Award, 2× Demo Award
- Outstanding Computer Science Student Award, Gettysburg College, 2013