https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/4-launch-agent/harness-run/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/task-walkthrough/https://dextermayhewjd.github.io/projects/slime/code-reading/https://dextermayhewjd.github.io/projects/slime/reuse-roadmap/https://dextermayhewjd.github.io/projects/python-async-concurrency/01-async-def/https://dextermayhewjd.github.io/courses/mit-1806/01-systems/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-1-slimes-core-components-and-architecture/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-1-slimes-core-components-and-architecture/1-1-1-high-performance-training-with-megatron-lm/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-2-extensibility-and-customization-points/1-2-1-dynamic-filtering-and-processing-hooks/https://dextermayhewjd.github.io/projects/slime/2-setting-up-your-development-environment/2-1-docker-image-construction-and-release-management/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/3-1-integrating-custom-dynamic-filtering-and-processing-hooks/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/https://dextermayhewjd.github.io/projects/slime/5-distributed-training-and-evaluation/5-1-configuration-of-distributed-training-and-evaluation-jobs/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/6-1-megatron-lm-compatible-model-implementations/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/6-1-megatron-lm-compatible-model-implementations/6-1-1-multi-latent-attention-mla-in-glm-5/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-1-advanced-coding-agent-rl-workflows/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-1-pytorch-profiler-trace-analysis-and-visualization/https://dextermayhewjd.github.io/projects/learn-claude-code/https://dextermayhewjd.github.io/courses/mit-1806/https://dextermayhewjd.github.io/projects/learn-claude-code/s01/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/01-structure/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/01-gotchas/https://dextermayhewjd.github.io/projects/learn-claude-code/s13/01-team-messaging/https://dextermayhewjd.github.io/projects/learn-claude-code/s15/system-prompt/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/syntax/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/2-translate-forward/data-flow/https://dextermayhewjd.github.io/projects/python-async-concurrency/https://dextermayhewjd.github.io/courses/cs285/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/courses/cs285/lecture-02/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/02-await/https://dextermayhewjd.github.io/courses/mit-1806/02-elimination/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-1-slimes-core-components-and-architecture/1-1-2-efficient-data-generation-and-rollouts-with-sglang/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-2-extensibility-and-customization-points/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-2-extensibility-and-customization-points/1-2-2-custom-data-generation-functions/https://dextermayhewjd.github.io/projects/slime/2-setting-up-your-development-environment/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/2-translate-forward/https://dextermayhewjd.github.io/projects/slime/2-setting-up-your-development-environment/2-2-npu-environment-setup-and-patching/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/3-2-configuring-and-validating-evaluation-datasets/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-2-backend-orchestration-for-megatron-lm-training/https://dextermayhewjd.github.io/projects/slime/5-distributed-training-and-evaluation/5-2-low-precision-training-and-inference/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/6-2-asynchronous-agent-trajectory-generation-rollout-buffer/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-2-multi-turn-vlm-training-with-interactive-environments/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-2-interactive-timeline-viewer-for-rollout-debug-dumps/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/asyncio/https://dextermayhewjd.github.io/projects/learn-claude-code/s02/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/02-real-expertise/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/02-output-templates/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/02-discovery/https://dextermayhewjd.github.io/projects/learn-claude-code/s13/02-team-protocols/https://dextermayhewjd.github.io/projects/learn-claude-code/s15/error-recovery/https://dextermayhewjd.github.io/projects/slime/https://dextermayhewjd.github.io/projects/python-async-concurrency/03-tasks/https://dextermayhewjd.github.io/courses/mit-1806/03-vector-spaces/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-2-extensibility-and-customization-points/1-2-3-custom-reward-functions/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-3-distributed-job-orchestration-and-management/https://dextermayhewjd.github.io/projects/slime/2-setting-up-your-development-environment/2-3-gb10-architecture-porting-and-optimization/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/3-generate-tokens/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/3-3-implementing-and-integrating-custom-reward-functions/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-3-sglang-backend-for-inference-serving-and-rollouts/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/6-3-model-conversion-utilities/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-3-on-policy-distillation-opd-for-model-alignment/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-3-advanced-quantization-strategies-fp8-int4-gptq-int4-direct-packing/https://dextermayhewjd.github.io/projects/learn-claude-code/s03/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/03-content-selection/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/03-checklists/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/03-invocation/https://dextermayhewjd.github.io/projects/learn-claude-code/s13/03-task-claiming/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/threading/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/3-5/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/pre-4/https://dextermayhewjd.github.io/projects/python-async-concurrency/04-results/https://dextermayhewjd.github.io/projects/slime/1-introduction-to-slime-framework/1-2-extensibility-and-customization-points/1-2-4-evaluation-dataset-configuration/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/3-4-developing-custom-data-generation-rollout-functions/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/4-launch-agent/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-4-distributed-ray-orchestration-and-resource-management/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/6-4-numerical-alignment-and-optimized-kernels/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-4-tool-enabled-language-model-generation-with-retool/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-4-dequantization-processes-fp8-to-bf16-k2-thinking-int4-to-bf16/https://dextermayhewjd.github.io/projects/learn-claude-code/s04/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/04-task-scope/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/04-validation-loops/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/04-context/https://dextermayhewjd.github.io/projects/learn-claude-code/s13/04-worktree-isolation/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/queue/https://dextermayhewjd.github.io/courses/cs285/lecture-05/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/05-cancellation/https://dextermayhewjd.github.io/projects/slime/3-extending-and-customizing-slime/3-5-adding-and-integrating-tests-into-ci/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-5-sample-generation-and-reward-modeling-pipeline/https://dextermayhewjd.github.io/projects/slime/5-distributed-training-and-evaluation/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/5-manage-sandbox/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-5-multi-turn-conversational-agents-with-search-r1/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-5-megatron-lm-to-hugging-face-model-conversion-and-verification/https://dextermayhewjd.github.io/projects/learn-claude-code/s05/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/05-control-calibration/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/05-plan-validation/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/05-rendering/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/executors/https://dextermayhewjd.github.io/courses/cs285/lecture-06/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/06-semaphore/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-6-core-utilities-for-system-level-functionality/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/6-exec-commands/https://dextermayhewjd.github.io/projects/slime/6-model-implementations-and-integration/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-6-multi-agent-reinforcement-learning/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/8-6-hugging-face-to-megatron-lm-model-conversion-and-distributed-saving/https://dextermayhewjd.github.io/projects/learn-claude-code/s06/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-instruction-patterns/06-scripts/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/06-permissions/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/multiprocessing/https://dextermayhewjd.github.io/projects/python-async-concurrency/07-queues/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/7-make-api-calls/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-7-fully-asynchronous-rollout-generation/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/07-references/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/07-subagents/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/subprocess/https://dextermayhewjd.github.io/projects/python-async-concurrency/08-executors/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-8-delta-weight-synchronization-for-disaggregated-systems/https://dextermayhewjd.github.io/projects/slime/8-performance-analysis-and-model-conversion-tools/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/08-evaluation/https://dextermayhewjd.github.io/projects/learn-claude-code/s08/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/signal/https://dextermayhewjd.github.io/projects/slime/4-slime-core-components-and-orchestration/4-1-agent-rollout-adapters-and-harnesses/trajectory-management/https://dextermayhewjd.github.io/courses/cs285/lecture-09/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/09-thread-synchronization/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-9-rollout-correction-methods-for-offline-rl/https://dextermayhewjd.github.io/projects/slime/9-observability-and-monitoring/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/09-agent-loop/https://dextermayhewjd.github.io/projects/learn-claude-code/s09/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/contextvars/https://dextermayhewjd.github.io/courses/cs285/lecture-10/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/10-async-protocols/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-10-agentic-multi-turn-tool-use-with-tau-bench/https://dextermayhewjd.github.io/projects/learn-claude-code/s07/10-appendix/https://dextermayhewjd.github.io/projects/learn-claude-code/s10/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/contextlib-async/https://dextermayhewjd.github.io/projects/python-async-concurrency/11-contextvars/https://dextermayhewjd.github.io/projects/slime/7-practical-examples-and-use-cases/7-11-integration-of-strands-sglang-for-agentic-rl/https://dextermayhewjd.github.io/projects/learn-claude-code/s11/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/inspect-async/https://dextermayhewjd.github.io/projects/python-async-concurrency/12-processes/https://dextermayhewjd.github.io/projects/learn-claude-code/s12/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/async-builtins/https://dextermayhewjd.github.io/projects/python-async-concurrency/13-event-loop/https://dextermayhewjd.github.io/projects/learn-claude-code/s13/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/async-typing/https://dextermayhewjd.github.io/courses/cs285/lecture-14/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/projects/python-async-concurrency/14-slime-walkthrough/https://dextermayhewjd.github.io/projects/learn-claude-code/s14/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/time-sleep/https://dextermayhewjd.github.io/projects/python-async-concurrency/15-async-notification/https://dextermayhewjd.github.io/projects/learn-claude-code/s15/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/os-process/https://dextermayhewjd.github.io/projects/python-async-concurrency/16-thread-coordination/https://dextermayhewjd.github.io/projects/learn-claude-code/s16/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/fcntl/https://dextermayhewjd.github.io/projects/python-async-concurrency/17-executor-results/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/select/https://dextermayhewjd.github.io/projects/learn-claude-code/s17/https://dextermayhewjd.github.io/projects/python-async-concurrency/18-process-lifecycle/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/coroutine-helpers/https://dextermayhewjd.github.io/projects/python-async-concurrency/19-signals-locks/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/candidates/https://dextermayhewjd.github.io/projects/python-async-concurrency/20-async-streams/https://dextermayhewjd.github.io/projects/python-async-concurrency/21-async-inspection/https://dextermayhewjd.github.io/projects/python-async-concurrency/coverage/https://dextermayhewjd.github.io/projects/python-async-concurrency/usage-index/https://dextermayhewjd.github.io/projects/python-async-concurrency/references/https://dextermayhewjd.github.io/tags/actor-critic/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/behavioral-cloning/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/cs285/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/dagger/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/gae/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/grpo/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/importance-sampling/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/irl/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/language-models/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/natural-gradient/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/policy-gradients/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/pomdp/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/ppo/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/reinforce/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/reparameterization/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/rlhf/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/trpo/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/trust-region/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/courses/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/tags/%E6%A8%A1%E4%BB%BF%E5%AD%A6%E4%B9%A0/2026-10-07T00:00:00+00:00https://dextermayhewjd.github.io/categories/https://dextermayhewjd.github.io/reviews/https://dextermayhewjd.github.io/notes/https://dextermayhewjd.github.io/about/https://dextermayhewjd.github.io/archives/https://dextermayhewjd.github.io/search/https://dextermayhewjd.github.io/projects/