Lexxs Xander

I’m an Agricultural and Bioresources Engineering graduate with hands-on experience in AI data annotation, automated systems design, and data pipeline management. I’ve built Python scripts to clean real-world agricultural datasets and support model-quality improvements through consistent labeling.

In remote roles, I annotate and quality-review large-scale datasets—at Atlas Capture, I label first-person (egocentric) action boundaries, object interactions, and task sequences using detailed taxonomy guidelines. At iMerit, I’ve labelled and reviewed 2,000+ image pairs per month, created structured change-detection reports, and maintained high accuracy (including 95%+ task accuracy) while delivering strong QA outcomes.

I also focus heavily on review quality and throughput: I run peer-review QA on 500+ submissions per cycle, provide structured feedback, and help reduce team error rates (including ~18% reported reduction). I collaborate asynchronously with project leads to resolve edge cases and refine labeling taxonomies—improving throughput by ~12% over review cycles.

Beyond annotation, I bring practical domain and engineering context from my NCAM internship and independent projects. I coordinated 5,000+ cassava-farm image annotations in CVAT and Roboflow (improving dataset accuracy by 30% and cutting training cycle time by 20%), and I built a web-based decision support system using Python and AHP for farm machinery selection, plus supporting data cleaning/enrichment pipelines.