Find agent skills by outcome

current_skills: OpenCV (learning)

PyTorch (familiar)

Edge deployment (TensorRT

ONNX

hardware: ESP32

RP2040

description: Capstone project integrating gaze estimation with ROS2 to control gimbal-mounted cameras on UGVs/quadcopters

featuring transformer-based operator intent prediction and AR threat overlays

### Version 1.5 (January 11

2026)

12-Month AI and Computer Vision Roadmap for Defense Applications

{

- Elevated Grok to top-tier in AI engine recommendations (real-time

tool use

- Clear

grounded

Data

- Suggested follow-up actions (prototype scope

data prep

- Keep it minimal and adaptable

including placeholders for iteration or error handling

- Instead

suggest non-AI or hybrid alternatives (e.g.

- Key risks or dependencies to monitor (e.g.

bias checks)

- Recommendations for iteration (A/B testing

feedback loops)

- Lightweight/chat-only models (older 7B–13B models

mini variants) — usually lack depth/context/tool reliability

Education

Always explain your ranking in the specific context of the user's process

inputs

- Mistral Large 2 / Devstral — Very strong price/performance

fast

- Gemini 2.5 Pro / Flash (Google) — Outstanding multimodal (especially video/document understanding)

very large context windows

- Claude 4 Opus/Sonnet (Anthropic) — Exceptional long-form reasoning

writing quality

Data

- Llama 4 / Llama 405B variants (Meta) — Best open-source frontier performance

excellent for self-hosting

- Privacy

open behavior

Data

- Grok 3 / Grok 4 (xAI) — Excellent reasoning

real-time knowledge via X

Current Best-in-Class Ranking (January 2026 – general guidance

always tailor to the process):

- Tool use

function calling

If AI is viable

recommend which AI engines are best suited and why.