Trishita Acharjee

I've built AI agents and production-style Python systems that connect external tools, APIs, and data into useful client-facing workflows.

At Caze Labs, I built Cloud CostSense, a Python diagnostic tool that compares AWS, Azure, and GCP pricing and produces automated cost-analysis reports for pre-deployment infrastructure decisions. I independently integrated and debugged third-party pricing APIs, resolving authentication failures, rate limits, and inconsistent response schemas.

My projects include Athena, a multi-tool AI agent using the Anthropic Claude API to decide when to use weather, web search, and persistent-memory tools, with traceable live activity logs. I also built InsightSQL, which turns plain-English questions into SQL, executes read-only queries, and explains results clearly while applying three layers of query safety controls.

Earlier, at Majestic Infotech Solutions, I built and tuned a Python and Scikit-Learn loan prediction model, improving accuracy by 15%. I enjoy diagnosing integration and data-quality problems, moving quickly from rough ideas to working prototypes, and explaining technical findings to non-technical stakeholders.