Raza Malik · ML & Robotics Engineer

Vanderbilt University ('28), Computer Science & Biological Sciences. I ship models to physical machines.

• Project I: Autonomous Excavator VLA & Control (Actor Labs) · pi0.5 VLA, Domain VLM (34% to 81% success), detector-v3 multi-task grounding, VR teleop & LiDAR BEV.
• Project II: Field Data Engine & Codecs (Actor Labs) · In-cab sensor pods, autonomous R2 Upload Box kiosks (868+ GB/node), DCVC-RT & 16-bit depth compression, SO-100 / WX200 testbeds.
• Project III: lore · Git-Native AI Memory (git-lore-cli) · X25519 + AES-256-GCM encrypted Git ref sync (refs/ai/lore) and SQLite FTS5 MCP server.
• Project IV: Island Ecosystem & Quadruped RL · 15-DOF deer locomotion in NVIDIA Isaac Lab (2,500 GPU envs), Whittaker procedural biome synthesis, Blender to USD to ONNX pipeline.
• Project V: aniML Vision (animlvision.com) · 90.7% top-1 / 97.6% class accuracy wildlife CV ensemble (YOLOv8 + EfficientNet-B3 + ConvNeXt) trained on 130,000+ images.
• Project VI: Tanzania Wildlife & Ecology · 5,000+ Serengeti & Tarangire field observations, QGIS elephant corridor & Maasai conflict mapping, Kilimanjaro off-grid solar engineering.
• Project VII: Anthro · Emergent Society Sim · Zero-tech-tree C# .NET Standard 2.1 & Unity anthropological simulation with 20-world paired bootstrap demographic evaluation.