Mind the Machine is a space for thinking about artificial intelligence before it thinks too much—or too confidently.
I am a Business Intelligence (BI) Architect with 18+ years of experience delivering enterprise-scale BI and analytics systems across domains such as Finance, CRM, and Marketing.
My professional work spans analytics platforms, data warehousing, and decision systems — including designing, operating, and sustaining BI solutions across on-premise and cloud environments.
Alongside core BI and analytics work, I focus on applied AI and automation, including AI-assisted development and the use of LLMs in data workflows, metadata creation, data governance, and system behavior analysis.
My writing explores how intelligent systems behave under real-world constraints — particularly quiet failures and loss of internal evaluation, where systems appear to perform well while gradually drifting toward breakdown.
My essays have been published with Towards AI and reflect a practical interest in understanding limits, failure modes, and why intelligence alone is insufficient in deployed systems.
These ideas are written to invite discussion — especially with teams building, deploying, and operating real-world AI and analytics systems.
You can find me on LinkedIn, X or follow my writing here on Medium. https://www.linkedin.com/in/carijit/
I also conduct formal research on inference-time regulation, evaluative coherence, and reliability in deployed language models.
Selected technical papers currently under peer review (preprints publicly available via DOI):
– Inference-Time Commitment Shaping: A Framework for Quiet Failure Mitigation in LLM Systems —https://zenodo.org/records/18401074 – Control Probe: Inference-Time Commitment Control — https://doi.org/10.5281/zenodo.18352963 – Evaluative Coherence Regulation (ECR): An Inference-Time Stability Layer for Enterprise LLMs — https://doi.org/10.5281/zenodo.18353477
Additional published working paper:
– Need for More Academic R&D in India — https://doi.org/10.2139/ssrn.1097502
These papers are technical and independent of my public writing.
