Typescript, Python, LLMs, harness engineering, Claude Code, on-device deployment, agentic patterns, applied machine learning, self-hosting using vLLM, SGLang

Jax, Tensorflow, Deep Learning, Pytorch, Spark, memory efficiency, GPU optimization, DPO, ORPO, Ray, RLHF pipelines, LLM inference frameworks, Megatron, Apache Arrow, ppo, distributed training frameworks, transformer architectures, training or fine-tuning large models, quantization, TensorRT-LLM, FasterTransformer, DeepSpeed, ZeRO, vLLM, large-scale data processing, mixed precision, FSDP, diffusion models, multimodal models