Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI Engineer with 1.5+ years of experience building production-grade generative AI systems, including multi-agent orchestration (LangGraph/LangChain) and RAG pipelines using Azure OpenAI and Azure AI Search. Skilled in Python and FastAPI, with end-to-end ownership from prototype to monitored deployments. Delivered agentic workflows that reduced manual processing time and improved incident resolution, while implementing evaluation, guardrails, and prompt-injection defenses for enterprise use cases.…AI Engineer with 1.5+ years of experience building production-grade generative AI systems, including multi-agent orchestration (LangGraph/LangChain) and RAG pipelines using Azure OpenAI and Azure AI Search. Skilled in Python and FastAPI, with end-to-end ownership from prototype to monitored deployments. Delivered agentic workflows that reduced manual processing time and improved incident resolution, while implementing evaluation, guardrails, and prompt-injection defenses for enterprise use cases.

Prateeksha Patel

AI Engineer, Data Scientist, Software Engineer, +1





AI Engineer with 1.5+ years of experience building production-grade generative AI systems, including multi-agent orchestration (LangGraph/LangChain) and RAG pipelines using Azure OpenAI and Azure AI Search. Skilled in Python and FastAPI, with end-to-end ownership from prototype to monitored deployments. Delivered agentic workflows that reduced manual processing time and improved incident resolution, while implementing evaluation, guardrails, and prompt-injection defenses for enterprise use cases.…AI Engineer with 1.5+ years of experience building production-grade generative AI systems, including multi-agent orchestration (LangGraph/LangChain) and RAG pipelines using Azure OpenAI and Azure AI Search. Skilled in Python and FastAPI, with end-to-end ownership from prototype to monitored deployments. Delivered agentic workflows that reduced manual processing time and improved incident resolution, while implementing evaluation, guardrails, and prompt-injection defenses for enterprise use cases.

Available to hire

AI Engineer with 1.5+ years of experience building production-grade generative AI systems, including multi-agent orchestration (LangGraph/LangChain) and RAG pipelines using Azure OpenAI and Azure AI Search.

Skilled in Python and FastAPI, with end-to-end ownership from prototype to monitored deployments. Delivered agentic workflows that reduced manual processing time and improved incident resolution, while implementing evaluation, guardrails, and prompt-injection defenses for enterprise use cases.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Language

Work Experience

Associate Data Scientist at CloudNexus DigiWeb Services Pvt. Ltd.

February 1, 2025 - Present

Architected multi-agent orchestration pipelines using LangGraph and LangChain, reducing manual processing time by 40% for core enterprise operations. Engineered RAG pipelines with Azure OpenAI embeddings and Azure AI Search via Azure AI Foundry, achieving sub-second query latency for grounded LLM outputs. Monitored production systems using LangSmith and Azure Application Insights, catching an estimated 30% more issues and reducing mean time-to-resolution by 25%. Built evaluation datasets and ground-truth benchmarks for retrieval precision and response faithfulness, reducing hallucinations by 35% before production. Implemented prompt-injection defenses and structured output validation using JSON schema/guardrails for agentic pipelines handling sensitive enterprise data.

Artificial Intelligence Developer at AskGalore Digital India Pvt. Ltd.

August 1, 2024 - December 31, 2024

Developed and deployed ML models for anomaly detection and real-time control optimization, packaged as documented FastAPI services for client inference. Automated data preprocessing pipelines and applied feature engineering, improving signal quality by an estimated 20% and reducing manual preprocessing time by roughly 6 hours per week. Designed evaluation frameworks for anomaly detection models with labeled test sets and precision/recall tracking to validate performance before client delivery.