Looking for a job

Jennica David

@jennicadavid

Message

I build data-driven AI products using RAG, NLP, machine learning, and analytics.

India

Message

What I'm looking for

I'm seeking opportunities to apply my analytical and engineering skills to build scalable, real-world, data-driven products, particularly in AI, NLP, RAG, machine learning, and data analytics.

I've built production-oriented AI and analytics solutions through internships at AngelBot AI and CDAC. At AngelBot AI, I contributed to a RAG-based document Q&A product by improving prompt engineering, vector database workflows, API integration, and model-evaluation feedback.

At CDAC, I built a Hindi part-of-speech tagging pipeline using CRF sequence labelling and MuRIL, evaluating models with accuracy, precision, recall, and F1 score. My work spans Python, SQL, machine learning, NLP, RAG, FastAPI, Elasticsearch, MySQL, and Power BI.

I've built DocuIntellect, a multimodal RAG document workspace with hybrid search, cited answers, knowledge graphs, and document categorization. I've also delivered customer and vendor analytics projects, deployed vendor classification through Flask, and contributed to two published Indian patents in environmental DNA analysis and privacy-preserving gait attendance.

Experience

Work history, roles, and key accomplishments

AI Intern

AngelBot AI Private Ltd.

May 2026 - Jul 2026 (2 months)

Contributed to a RAG-based document Q&A product, working across prompt engineering, vector databases, and API integration to improve answer accuracy and grounding. Debugged issues across the RAG pipeline and evaluated model outputs, feeding structured feedback into the system to support model-evaluation cycles and feature development.

RAG Prompt Engineering Vector Databases API Integration Model Evaluation Debugging

NLP Intern

CDAC (Centre for Development of Advanced Computing)

Jun 2025 - Aug 2025 (2 months)

Built a Hindi Part-of-Speech tagging pipeline combining a CRF sequence-labelling model with the transformer-based MuRIL model on the BIS tagset. Evaluated both models on custom test sets using Accuracy, Precision, Recall, and F1 Score to guide model selection.