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

AI/ML Engineer with 3.5+ years of experience shipping production GenAI and ML systems for finance and tech, including multi-agent systems and RAG pipelines. Focused on evaluation that stands up to recall/faithfulness and compliance scrutiny, with explainability that risk and governance teams can defend. Hands-on in fine-tuning transformers (LoRA/QLoRA/PEFT), and deploying/monitoring models across AWS and Azure. Built knowledge graphs + RAG with automated evaluation harnesses, and delivered real-time analytics for executive decision-making in regulated environments.…AI/ML Engineer with 3.5+ years of experience shipping production GenAI and ML systems for finance and tech, including multi-agent systems and RAG pipelines. Focused on evaluation that stands up to recall/faithfulness and compliance scrutiny, with explainability that risk and governance teams can defend. Hands-on in fine-tuning transformers (LoRA/QLoRA/PEFT), and deploying/monitoring models across AWS and Azure. Built knowledge graphs + RAG with automated evaluation harnesses, and delivered real-time analytics for executive decision-making in regulated environments.

Arshita R

AI Engineer, Data Scientist, Programmer, +1





AI/ML Engineer with 3.5+ years of experience shipping production GenAI and ML systems for finance and tech, including multi-agent systems and RAG pipelines. Focused on evaluation that stands up to recall/faithfulness and compliance scrutiny, with explainability that risk and governance teams can defend. Hands-on in fine-tuning transformers (LoRA/QLoRA/PEFT), and deploying/monitoring models across AWS and Azure. Built knowledge graphs + RAG with automated evaluation harnesses, and delivered real-time analytics for executive decision-making in regulated environments.…AI/ML Engineer with 3.5+ years of experience shipping production GenAI and ML systems for finance and tech, including multi-agent systems and RAG pipelines. Focused on evaluation that stands up to recall/faithfulness and compliance scrutiny, with explainability that risk and governance teams can defend. Hands-on in fine-tuning transformers (LoRA/QLoRA/PEFT), and deploying/monitoring models across AWS and Azure. Built knowledge graphs + RAG with automated evaluation harnesses, and delivered real-time analytics for executive decision-making in regulated environments.

Available to hire

AI/ML Engineer with 3.5+ years of experience shipping production GenAI and ML systems for finance and tech, including multi-agent systems and RAG pipelines. Focused on evaluation that stands up to recall/faithfulness and compliance scrutiny, with explainability that risk and governance teams can defend.

Hands-on in fine-tuning transformers (LoRA/QLoRA/PEFT), and deploying/monitoring models across AWS and Azure. Built knowledge graphs + RAG with automated evaluation harnesses, and delivered real-time analytics for executive decision-making in regulated environments.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

AI Engineer at Wells Fargo

January 1, 2025 - Present

Built a multi-agent financial compliance auditing platform using CrewAI and LangGraph to automate regulatory risk assessment across institutional documents. Increased audit throughput by 34% and improved alert precision, cutting false positives by 22% while holding recall constant via evidence-grounded validation. Designed a Knowledge Graph + RAG pipeline (LlamaIndex, LLaMA, Pinecone) with an evaluation harness; achieved 94% retrieval recall@5 and 96% answer faithfulness while reducing retrieval latency by 180ms. Fine-tuned a BERT-family transformer with LoRA for NER on unstructured cross-border transaction logs, reaching 0.91 macro-F1 and reducing downstream processing time by 41%. Productionized a credit risk model on Azure AI Studio with LightGBM and TensorFlow behind a monitored inference endpoint with drift detection and scheduled retraining; improved AUC-ROC by ~14% and increased minority default recall.

ML Engineer at Tech Mahindra

August 1, 2021 - July 31, 2023

Fine-tuned pre-trained BERT models using LoRA/PEFT (Hugging Face) and NLTK for ticket-category classification, tracking macro-F1 in validation and reducing customer support ticket resolution time by 15 minutes. Deployed an enterprise demand forecasting system using ARIMA, Prophet, and XGBoost on AWS SageMaker, reducing forecast error (WMAPE) by 26% versus a seasonal-naive baseline and decreasing stockouts. Built a defect detection pipeline for manufacturing clients using Mask R-CNN and PyTorch, accelerating QA inspection by 38% while maintaining 94% precision and strong recall on structural anomalies. Implemented client segmentation using PCA + K-Means/DBSCAN on PySpark across 2M+ user profiles with silhouette and business KPI validation. Engineered scalable ETL pipelines in Python and PostgreSQL to clean/process unstructured logs and stream validated features into Tableau dashboards for executive performance tracking.