Nihal Pardeshi – Product-Oriented AI Builder

Summary

I design and deploy GenAI systems using structured prompt engineering, retrieval pipelines, and multi-step reasoning workflows that connect speech, context, and action.

Model capability is only one part of the system. The real challenge is defining when the model should act, how it should reason, and how it should fail safely.

I don't treat AI as a feature layer. I treat it as a decision system that must be scoped, controlled, and evaluated.

Approach

Discovery

Translate ambiguous operational problems into structured product requirements.

Specification

Define LLM flows, retrieval logic, fallback paths, and edge-case handling before deployment.

Evaluation

Measure across latency thresholds, completion reliability, failure mode frequency, and manual intervention rate.

Iteration

Refine prompts, context windows, and orchestration layers based on observed behavior — not intuition.

Experience

AI Intern – ChatMaven.ai ↗

Jun 2025 – Nov 2025

+Defined and launched an AI-powered healthcare IVR by translating operational pain points into prioritized automation workflows aligned with business outcomes.
+Drove requirement definition and feature prioritization with engineering and stakeholders, ensuring alignment between technical feasibility and user impact.
+Optimized LLM-driven voice, SMS, and email workflows to improve system reliability and reduce manual intervention in live customer operations.
+Owned end-to-end delivery from problem definition to AWS deployment, balancing iteration speed with production stability.