Looking for a job

Smith Pateliya

@smithpateliya

AI/ML engineer at Intuz building production LLM and document intelligence backends, including ESG retrieval and auditing workflows.

India

Message

What I'm looking for

I'm interested in AI/ML engineering and Python backend opportunities involving LLM systems, retrieval, document intelligence, and practical model deployment.

At Intuz, I build production AI backends for LLM orchestration, document intelligence, computer vision, and on-device inference. I architected ESRS sustainability auditing workflows using LangGraph, fine-tuned Mistral 7B, Claude Sonnet, and Milvus.

I also built a multi-report ESG retrieval backend with Qdrant, Cohere embeddings and reranking, and AWS Textract. It uses KPI-aware query rewriting and persistent conversational context.

For offline Android speech-to-text, I delivered a Whisper.cpp workflow with React Native and a Kotlin/C++ JNI bridge. I also trained a YOLO pipeline to identify elements in engineering drawings and connect them to contextual analysis and vector-based retrieval.

Earlier, at Career-Naksha, I built a pipeline that transformed state budget PDFs into structured datasets. As a Data Scientist Intern at Evoastra Ventures Pvt. Ltd., I built automotive market data ingestion and analysis workflows; I’m interested in AI/ML engineering and Python backend opportunities focused on LLM systems, retrieval, document intelligence, and practical model deployment.

Experience

Work history, roles, and key accomplishments

Current

AI / ML Engineer | Jr. Software Engineer

Current

Intuz

Jan 2025 - Present (1 year 10 months)

I build production AI backends across LLM orchestration, document intelligence, computer vision, and on-device inference.
• Architected ESRS sustainability auditing workflows with LangGraph, fine-tuned Mistral 7B, Claude Sonnet, and Milvus retrieval for evidence checks, clarification, proofreading, and report comparisons.
• Built a multi-report ESG RAG backend using Qdrant, Cohere Embed v4.0, Rera

Data Analyst Intern

Career-Naksha

Sep 2024 - Nov 2024 (2 months)

• Built an automated pipeline to transform unstructured state budget PDFs into structured datasets through parsing, validation, and schema normalization.
• Delivered analytical outputs published on careernaksha.com to support career-counselling recommendations.
Skills: Python, PDF extraction, data cleaning, data validation, ETL, data analysis.

Data Scientist Intern

Evoastra Ventures Pvt. Ltd.

Jul 2024 - Sep 2024 (2 months)

• Built a Selenium and BeautifulSoup scraping backend to ingest automotive market data with automated scheduling and fault-tolerant retries.
• Developed reusable data cleaning, feature engineering, and exploratory analysis workflows to support pricing intelligence by fuel category and transmission type.
Skills: Python, Selenium, BeautifulSoup, Pandas, data cleaning, feature engineering, explorator