Graduate researcher at Missouri S&T building AI tools for structural health monitoring, autonomous robotics, and geospatial deep learning. Founder of Meridian.
I'm a Computer Science M.S. student at Missouri University of Science and Technology, where my research focuses on AI-driven frameworks for crack width measurement in structural health monitoring imagery. My thesis develops annotation pipelines, segmentation models, and width estimation algorithms to automate structural inspection.
Previously a USGS student contractor co-authoring peer-reviewed work on deep learning for hydrographic feature extraction, and a UF Pathfinder Fellow working on 3D reconstruction and autonomous navigation for greenhouse robotics.
I also built and ship Meridian, a production MCP coordination server for parallel AI coding sessions.
Production MCP server for coordinating parallel AI coding sessions. Multi-tenant SaaS with per-tenant Neon DB isolation, AES-256 encryption, OAuth + magic link auth, Stripe billing, 535+ tests, CI/CD, and PyInstaller cross-platform binaries. 54+ coordination tools. Built and dogfooded using its own sprint-tracking infrastructure.
M.S. thesis. PyQt5 annotator and evaluation pipeline for crack midline extraction and sub-pixel width estimation. Multi-cue approach (segmentation + depth + intensity + RGB) recovers true geometric width perpendicular to the crack centerline — beating binary mask baselines. Benchmarked against SOTA on SUT-Crack dataset.
Detects blueberry flowers in field imagery and classifies their phenological development stage. Supports precision horticulture timing and yield prediction workflows.
Team lead. Built a PyQt5 GUI tool for semi-automatic video annotation using SAM3.1 — generates, modifies, and propagates first-frame segmentation masks through video for YOLO training data generation. Used for bamboo culm segmentation in a robotics perception project at UF.
Pathfinder Fellow project at University of Florida. Compared monocular AI reconstruction (VGGT, Pi3) to traditional SLAM/TSDF (Cupoch/Open3D, NVBlox, GPS SLAM) on Azure Kinect RGB-D data in a vanilla greenhouse. Also prototyped FreeCAD pollination end effector and Isaac Sim airflow simulation.
2-year student contractor role. Built arcpy/geopandas workflows for the National Map, automated 10TB AK HUC12 data processing for CNN inputs, D-infinity augmentation for connectivity improvement, S3 sinuosity metric animations with audified piano score. 6 co-authored publications.
Summer internship at Washington University School of Medicine. Analyzed a pilot study of novel cancer detection via circRNA in urine across 110 renal cancer, bladder cancer, and healthy samples. Discovered significant differential gene expression. Built pipelines using CircExplorer, STAR, featureCounts, edgeR, and EDA between linear and circular RNA.
Capstone project for startup MARSfarm. Team of 3. Improved plant masking with custom OpenCV algorithms vs PlantCV baseline. Pioneered a PlantCV feature for automatic ROI label generation (merged to main). Deployed scalable pixel count algorithm on EC2 for 230k images on S3.
Closed-loop temperature and humidity controller on Raspberry Pi (ABE5009 — Control Systems in SmartAg). DHT11 sensor + MOSFET-switched heater and fan. Implemented PID vs Fuzzy-PID: PID achieved 0.17°C steady-state error; Fuzzy-PID achieved 3.6× faster rise time (51s vs 187s). Simulink plant model + real-time matplotlib dashboard.
3-node XBee wireless sensor network (ABE6031). Arduino Uno (DHT11 + SGP30 eCO₂/TVOC + photoresistor) + Elegoo MEGA 2560 (DHT11 + photoresistor). Raspberry Pi 4 coordinator parses serial data from all nodes, logs to CSV, uploads 8-channel telemetry to ThingSpeak every 20s, captures Pi Camera images to Google Drive via rclone every 60s.
Five-stage sensor signal processing pipeline in LabVIEW (ABE6031 Take-Home Final): (1) Z-score outlier removal (σ=2), (2) IIR Butterworth lowpass at 400Hz, (3) FFT with Hanning window + peak magnitude measurement, (4) Fourier series reconstruction at 4 dominant frequencies (~100, 200, 300, 401Hz), (5) RMS error = 4.84 between filtered and reconstructed signals.
Graduate ML project (CS6001, w/ Luke Hendrickson) fine-tuning DNABERT-2 for per-base DNA sequencing error detection. 3.2M 150bp reads from chr21 T2T reference with 10% indel + 1% substitution noise. Applied LoRA, fixed Triton Flash Attention bug. Trained on A100 80GB. F1: 0.651 → 0.995; Recall: 0.486 → 1.0.
Godot-based interactive logic circuit builder for ECE education. Staff can create and save customizable quiz layouts. Features 12 drag-and-drop circuit components with full connection logic.
As EECS Head of Department for Missouri S&T's Sustainable Engineering Design Team, drafted the Vertical Farming Feasibility Study, developed a Docker + Grafana greenhouse dashboard on a Mini PC with ESP32 sensors, built the team's WordPress website, and ran IoT/backend tutorials for team members over multiple semesters.
@inproceedings{stanislawski2025density,
author = {Stanislawski, Lawrence and Walker, B.A. and Kronenfeld, B.J. and Buttenfield, B.P. and Shavers, E.J. and Camerer, Adam and Patel, Y.},
title = {Assessment of Density Pattern Retention of Generalized Data for 1:100,000-scale United States Topographic Maps},
booktitle = {ICA Workshop on Multiscale Cartography and Sustainability},
address = {Vancouver, Canada},
year = {2025}
}
@inproceedings{stanislawski2025mapping,
author = {Stanislawski, Lawrence V. and Shavers, Ethan and Pastick, Neal and Thiem, Peter and Wang, Shaowen and Jaroenchai, Nattapon and Camerer, Adam and Kronenfeld, Barry and Buttenfield, Barbara P.},
title = {Mapping and Monitoring Surface Water Features in {Alaska} through Machine Learning},
booktitle = {AWRA Spring Conference},
address = {Anchorage, AK},
year = {2025}
}
@inproceedings{stanislawski2024adaptive,
author = {Stanislawski, Lawrence V. and Shavers, Ethan and Pastick, Neal and Thiem, Peter and Wang, Shaowen and Jaroenchai, Nattapon and Jiang, Zhe and Kronenfeld, Barry and Buttenfield, Barbara P. and Camerer, Adam},
title = {Adaptive Fine-tuning for Transferring a {U-net} Hydrography Extraction Model using {K-means}},
booktitle = {CaGIS/UCGIS 2024 Symposium},
address = {Columbus, OH},
year = {2024}
}
@inproceedings{stanislawski2024transfer,
author = {Stanislawski, Lawrence and Shavers, Ethan and Pastick, Neal and Thiem, Peter and Wang, Shaowen and Jaroenchai, Nattapon and Jiang, Zhe and Camerer, Adam and Kronenfeld, Barry and Buttenfield, Barbara P.},
title = {Implementation of Transfer Learning to Extract Hydrography for {Alaska} from {IfSAR} Data},
booktitle = {AWRA/UCOWR/NIWR Joint Water Resources Conference},
address = {St. Louis, MO},
year = {2024}
}
@inproceedings{stanislawski2024assessment,
author = {Stanislawski, Lawrence V. and Kronenfeld, Barry J. and Buttenfield, Barbara P. and Shavers, Ethan and Camerer, Adam},
title = {Assessment of Density Pattern Retention for Generalized Data},
booktitle = {AAG 2024 Annual Meeting},
address = {Honolulu, HI},
year = {2024}
}
@inproceedings{stanislawski2023transferring,
author = {Stanislawski, Lawrence V. and Jaroenchai, Nattapon and Wang, Shaowen and Shavers, Ethan and Duffy, Alex and Thiem, Peter and Jiang, Zhe and Camerer, Adam},
title = {Transferring Deep Learning Models for Hydrographic Feature Extraction from {IfSAR} Data in {Alaska}},
booktitle = {{GIScience} 2023 Workshop on {CartoAI}: {AI} for Cartography},
address = {Leeds, UK},
year = {2023},
url = {https://pubs.usgs.gov/publication/70250212}
}
3D reconstruction of vanilla greenhouse with monocular-camera AI (VGGT, Pi3) vs traditional SLAM/TSDF (Cupoch/Open3D, NVBlox, GPS SLAM) for Azure Kinect RGB-D camera. Prototyped FreeCAD vanilla pollination end effector and Unreal/Isaac Sim airflow simulation.
GitHubBuilt a unified crack midline and width evaluation PyQt5 annotator/pipeline comparing baseline segmentation (SAM3, HrSegNet, CrackScopeNet) and width estimation methods for complex crack geometry. Multi-cue centerline extraction pipeline; transfer learning experiments for HrSegNet raising crack IoU from ~0% to 7.49% on a challenging benchmark.
GitHubDeveloped arcpy and geopandas workflows for the National Map, automated 10TB AK HUC12 data processing for CNN inputs, D-infinity augmentation for connectivity improvement, S3 sinuosity metric animations with audified piano score. 6 co-authored publications.
USGS PublicationAnalyzed pilot study of novel cancer detection via circRNA in urine for 110 samples. Discovered significant differences in cancer vs healthy genes. Built pipelines for CircExplorer, featureCounts, edgeR analysis, and EDA between linear and circular RNA.
GitHubAssisted in creation of smart factory of Additive Manufacturing CNC machines. Prototyped IoT data pipeline to AWS (S3, DynamoDB, Analytics, CloudWatch). Built Cloudflare + Node-RED server on Raspberry Pi for remote sensor monitoring.
GitHubDrafted Vertical Farming Feasibility Study. Built team WordPress website. Prototyped Docker + Grafana greenhouse dashboard using Mini PC and ESP32. Ran IoT and backend tutorials for team members across multiple semesters.
Org GitHub Website