AI & Automation Consultant | Data Scientist | ML Engineer
Houston, TX
LinkedIn Β· Portfolio Β· AI Chronicle Β· satishkc19@gmail.com
I architect production AI systems that turn complex data problems into measurable business outcomes - spanning ML pipelines, LLM-powered automation, and end-to-end system design.
Currently at Bridgital - building distributed automation systems, generative AI agents with compliance guardrails, and semantic search pipelines on AWS Bedrock.
Previously - OCR and document classification systems at Bitpoint (75% reduction in manual entry across 500+ daily medical bills) and a multi-domain legal RAG system at Hartsfield & Associates (91% Recall@5 across 1,500+ documents).
| Project | Contribution | PR |
|---|---|---|
openclaw/openclaw |
Add AllowTcpForwarding prerequisite to Hetzner SSH tunnel docs | #54564 |
langgenius/dify |
Refactor: replace dict params with BaseModel in AppService | #35904 |
ggml-org/llama.cpp |
Fix WebUI CORS proxy missing API key for MCP connections | #21193 |
pytorch/pytorch |
Fix misleading out_dtype signatures in bmm/mm/addmm docs |
#179182 |
keras-team/keras |
Fix DepthwiseConv1D channels_first dilation bug |
#22641 |
roboflow/supervision |
Fix: preserve audio stream in process_video |
#2252 |
deepset-ai/haystack |
Update llama.cpp repository links to ggml-org | #10964 |
optuna/optuna |
Move type-only imports into TYPE_CHECKING block |
#6609 |
online-ml/river |
Per-sample weights in progressive validation | #1795 |
qutip/qutip-qip |
Default zero state for QubitCircuit.run() |
#376 |
- AWS Certified AI Practitioner - Amazon Web Services (Verify)
- MLOps: Getting Started - Google Cloud
- Transformer Models and BERT - Google Cloud
- Encoder-Decoder Architecture - Google Cloud
- AI Demystified - IBM WatsonX
Case studies and project breakdowns covering production ML systems, LLM pipelines, and automation infrastructure.
Weekly research publication breaking down the latest AI papers and automation frameworks for practitioners. Covers LLMs, agents, deep learning, quantization, and production deployment patterns β with visual summaries, technical depth, and estimated read times. Includes weekly Zoom discussions for subscribers.



