A thousand tools, one small backpack
Put Codex, Claude Code, Hermes Agent, and DeepSeek-style Code Mode beside the same hypothetical 50,000-tool company and compare what actually reaches the model.
kaushal@kforcode:~$ whoami
Engineering AI systems built for production.
I’m Kaushal, a Staff AI/ML Engineer. I build agent runtimes, evaluation systems, retrieval infrastructure, and developer platforms—and write about what actually works.
Building at HighLevel8+ years in applied AI3 peer-reviewed papers
FIG.01 · the_system_layer
latest_writingview archive →
→
Open model accuracy is a harness problem
Open model accuracy is also a systems problem. Hermes Agent narrows tool choices, repairs safe mistakes, preserves context, and helps smaller models finish.
hermes-agent architecture: caching outranks compaction
The hermes-agent architecture at a glance: why the largest open-source agent harness ships both of its cheap context reducers switched off, and what that buys.
hermes-agent internals, one component at a time
A component-by-component study of hermes-agent internals: the loop, the prompt, the cache plan, compaction, state, approvals, and what each one costs.
working_set
Systems thinking for applied AI.
I write for engineers, researchers, and technical leaders working where model behavior meets reliability, latency, cost, and product constraints.
15M+conversations / quarter
800B+tokens / quarter
600+MCP operations
99.9%runtime success
selected_researchall papers →
Research informed by real systems.
Peer-reviewed work on Document AI, privacy, clinical summarization, and efficient fine-tuning.
- 01 · 2024De-Identification of Sensitive Personal Data in IIT-CDIP-Derived DatasetsEMNLP 2024 (Main) ↗
- 02 · 2024Parameter-Efficient Fine-Tuning for Hospital Discharge SummarizationElsevier — Applied Soft Computing ↗
- 03 · 2025DoRA+: Enhancing Weight-Decomposed Low-Rank AdaptationElsevier — Applied Soft Computing ↗
open_channel
FOR::TEAMS
Building AI products?
I consult on agent platforms, retrieval quality, evals, and LLM cost — advisory sessions, architecture reviews, and ongoing engagements.
work with me →FOR::RECRUITERSHiring Staff+ AI engineers?
Eight years of production AI across four companies, three peer-reviewed papers, and the platform behind six AI product surfaces.
see the track record →