Open to opportunities

Swati Chaudhari

@swatichaudhari

Senior AI/ML engineer building production LLM systems: agentic RAG, multimodal document intelligence, and responsible guardrails.

India

Message

What I'm looking for

I’m looking to lead production LLM/agentic RAG systems—agentic orchestration, multimodal document intelligence, and robust responsible-AI with full observability—on a team that values quality, safety, and measurable impact (remote/relocation OK).

I’m a Senior AI/ML engineer specializing in production LLM systems—agentic RAG, multi-agent orchestration, and multimodal document intelligence at enterprise scale. I architect end-to-end solutions across prompt engineering, evaluation, responsible-AI guardrails, and full production observability.

I’ve architected and shipped 6+ end-to-end systems, including agentic RAG, multimodal document intelligence, NL-to-SQL, voice AI automation, and ML recommendations. I also build engineering standards like RAGAS-style evaluation pipelines, structured output enforcement, and secure API design with 85%+ test coverage—so quality is measurable, not assumed.

In my current role as Senior AI Engineer (AI Systems Lead) at Plansquare, I lead delivery of a production agentic RAG platform for enterprise document intelligence. I build multimodal pipelines that handle text, tables, and embedded images using layout parsing approaches (e.g., LayoutLM/Donut/LLaVA), and I deliver an enterprise NL-to-SQL platform with access control, query safety validation, and analytics reporting.

Previously, I built a production voice AI automation platform with multi-agent intent routing and improved customer-facing consistency through few-shot prompt iteration. I also drove 30% engagement gains and 20% operational efficiency improvements in multimodal AI features at Fitelo, and I bring 7 years of teaching and PhD research momentum into how I mentor teams and translate business goals into reliable AI systems.

Experience

Work history, roles, and key accomplishments

Current

Senior AI Engineer

Current

Plansquare Procon Pvt Ltd

Jun 2025 - Present (11 months)

Led architecture and delivery of a production agentic RAG platform for enterprise document intelligence, including multi-stage hybrid retrieval and context-window management. Built multi-agent orchestration workflows, multimodal document intelligence pipelines, and an NL-to-SQL platform with responsible-AI guardrails and 85%+ test coverage.

Agentic RAG LangGraph Layout Parsing (LayoutLM Donut LLaVA)NL To SQL RBAC RAGAS Style Evaluation

Senior Data Scientist

Astria Digital

Sep 2024 - May 2025 (8 months)

Built a production voice AI automation platform with multi-agent intent routing and personalized recommendation workflows. Designed and iterated few-shot prompt pipelines with structured output enforcement to reduce failure modes and improve response consistency for customer-facing interactions.

Multi Agent Intent Routing Customer Facing LLM Reliability Model Evaluation Cross Functional Architecture Translation

Data Scientist

Fitelo

Jan 2024 - Sep 2024 (8 months)

Developed multimodal AI features combining text and image understanding for personalized wellness recommendations at scale. Implemented quality checkpoints with human-in-the-loop feedback (30% engagement increase, 20% operational efficiency improvement) and maintained structured output validation with responsible-AI safety guardrails.

Multimodal AI Text & Image Understanding Human In The Loop Quality Gates Recommendation Systems Pipeline Design

Machine Learning Engineer

Plansquare Procon Pvt Ltd

Oct 2022 - Nov 2023 (1 year 1 month)

Designed and deployed a hybrid NLP recommendation engine combining collaborative filtering with content-based signals. Built and maintained end-to-end ML pipelines (preprocessing, feature engineering, training, evaluation, and deployment), delivering 20% engagement growth, 15% retention improvement, and 12% revenue lift.

Hybrid Recommendation Engines Collaborative Filtering Content Based NLP Signals Feature Engineering data preprocessing Experimentation