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

I'm Pratyush Joshi, an AI Engineer focused on building reliable, scalable AI systems—especially around LLMs, multi-agent orchestration, RAG pipelines, and real-time decisioning for finance and risk. I enjoy turning complex data into practical risk insights and collaborating with product, risk, and compliance teams to ship responsible AI. I design cloud-native, production-grade ML platforms with strong emphasis on model observability, explainability, and context-aware outputs. My experience spans fraud detection, credit risk modelling, and quantitative risk analytics within banking and financial services, where I thrive on solving hard problems with robust, auditable solutions.…I'm Pratyush Joshi, an AI Engineer focused on building reliable, scalable AI systems—especially around LLMs, multi-agent orchestration, RAG pipelines, and real-time decisioning for finance and risk. I enjoy turning complex data into practical risk insights and collaborating with product, risk, and compliance teams to ship responsible AI. I design cloud-native, production-grade ML platforms with strong emphasis on model observability, explainability, and context-aware outputs. My experience spans fraud detection, credit risk modelling, and quantitative risk analytics within banking and financial services, where I thrive on solving hard problems with robust, auditable solutions.

Pratyush Joshi





I'm Pratyush Joshi, an AI Engineer focused on building reliable, scalable AI systems—especially around LLMs, multi-agent orchestration, RAG pipelines, and real-time decisioning for finance and risk. I enjoy turning complex data into practical risk insights and collaborating with product, risk, and compliance teams to ship responsible AI. I design cloud-native, production-grade ML platforms with strong emphasis on model observability, explainability, and context-aware outputs. My experience spans fraud detection, credit risk modelling, and quantitative risk analytics within banking and financial services, where I thrive on solving hard problems with robust, auditable solutions.…I'm Pratyush Joshi, an AI Engineer focused on building reliable, scalable AI systems—especially around LLMs, multi-agent orchestration, RAG pipelines, and real-time decisioning for finance and risk. I enjoy turning complex data into practical risk insights and collaborating with product, risk, and compliance teams to ship responsible AI. I design cloud-native, production-grade ML platforms with strong emphasis on model observability, explainability, and context-aware outputs. My experience spans fraud detection, credit risk modelling, and quantitative risk analytics within banking and financial services, where I thrive on solving hard problems with robust, auditable solutions.

Available to hire

I’m Pratyush Joshi, an AI Engineer focused on building reliable, scalable AI systems—especially around LLMs, multi-agent orchestration, RAG pipelines, and real-time decisioning for finance and risk. I enjoy turning complex data into practical risk insights and collaborating with product, risk, and compliance teams to ship responsible AI.

I design cloud-native, production-grade ML platforms with strong emphasis on model observability, explainability, and context-aware outputs. My experience spans fraud detection, credit risk modelling, and quantitative risk analytics within banking and financial services, where I thrive on solving hard problems with robust, auditable solutions.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Language

Work Experience

Data Scientist / ML Engineer at Independent Project - Food Waste Optimisation Platform

August 1, 2025 - Present

Designed a data-driven system to model restaurant food consumption and per-item demand using time-series signals and contextual variables (customer flow, weather, events). Built predictive models for per-customer demand with lag features and rolling statistics, and implemented optimization logic to generate next-day preparation recommendations balancing waste reduction, stockout risk, and inventory constraints. Created end-to-end pipelines (POS data, inventory, recipes) and scalable APIs for real-time inference and daily planning, underpinned by a dynamic, stochastic system model to improve decision-making.

Associate Software Engineer - Data Science (Insurance Fraud) at Ernst & Young GDS

March 1, 2022 - May 1, 2023

Built ML models to detect insurance fraud, improving detection accuracy by 30% and reducing false positives by 15%, helping prevent substantial fraudulent payouts annually. Developed scalable PySpark ETL pipelines for near real-time fraud monitoring, and automated preprocessing and model training workflows to save manual effort and improve productivity. Conducted A/B tests to select top-performing models and integrated dashboards with cross-functional teams to accelerate investigation and claims efficiency.

Programmer Analyst Trainee - Data Science (Banking) at Cognizant Technology Solutions

January 1, 2021 - March 1, 2022

Developed credit risk ML models improving loan default prediction. Implemented drift-detection and monitoring pipelines for live models, enabling proactive risk management. Automated reporting and KPI monitoring, transitioned batch scoring to near real-time, and ensured fairness and regulatory compliance to mitigate penalties.