Looking for a job

Anshul Prajapati

@anshulprajapati

Message

AI/ML Engineer at Intuit building grounded financial RAG workflows and improving retrieval precision@5 by 13%.

Canada

Message

At Intuit, I build RAG pipelines for bookkeeping and financial-assistant use cases, grounding answers in approved tax and accounting content. I improved precision@5 by 13% on a financial-query benchmark and helped reduce unsupported answers by 20% on the offline test set.

At AMD, I benchmarked PyTorch transformer inference and built a configurable framework that evaluated 40 model setups. At Zoho, I built CRM lead-scoring and deal-win prediction models, and automated a Python/SQL batch-scoring pipeline for CRM activity records.

Experience

Work history, roles, and key accomplishments

Current

AI/ML Engineer

Current

Intuit

Jun 2026 - Present (5 months)

Build RAG pipelines for bookkeeping and financial-assistant use cases, integrating LLM APIs into Python backends with guardrails. Improved precision@5 by 13% and reduced p95 latency by 18% on staging.

Python SQL RAG LLM Integrations Prompt Engineering GuardRails PII Detection Prompt Injection Testing Async Processing Model Evaluation Precision@K Latency Optimization

AI/ML Intern

AMD

Jan 2026 - Apr 2026 (3 months)

Benchmarked PyTorch inference for transformer models, reducing median latency by 16% and peak memory usage by 12%. Built a config-driven evaluation framework and an LLM-based log summarizer.

PyTorch Transformer Models LLM Benchmarking Python Model Evaluation Memory Optimization Debugging

AI/ML Engineer (Contract)

Zoho

Jan 2022 - Dec 2024 (2 years 11 months)

Built lead-scoring and deal-win prediction models with XGBoost and scikit-learn, improving top-decile precision by 11%. Automated batch-scoring pipelines and shipped models as REST APIs in Docker.