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

I’m Prakhar Mehta, an AI/ML engineer focused on shipping reliable machine learning systems into production. With 4+ years of experience, I build end-to-end pipelines, develop models, and craft API-based inference services that scale, perform, and stay maintainable. I have hands-on experience with LLMs, retrieval-based systems, and real-time prediction services, and I enjoy collaborating with cross-functional teams to integrate AI features with existing applications and business goals.…I’m Prakhar Mehta, an AI/ML engineer focused on shipping reliable machine learning systems into production. With 4+ years of experience, I build end-to-end pipelines, develop models, and craft API-based inference services that scale, perform, and stay maintainable. I have hands-on experience with LLMs, retrieval-based systems, and real-time prediction services, and I enjoy collaborating with cross-functional teams to integrate AI features with existing applications and business goals.

Prakhar Mehta

AI Engineer, Data Scientist, AI Developer, +2





I’m Prakhar Mehta, an AI/ML engineer focused on shipping reliable machine learning systems into production. With 4+ years of experience, I build end-to-end pipelines, develop models, and craft API-based inference services that scale, perform, and stay maintainable. I have hands-on experience with LLMs, retrieval-based systems, and real-time prediction services, and I enjoy collaborating with cross-functional teams to integrate AI features with existing applications and business goals.…I’m Prakhar Mehta, an AI/ML engineer focused on shipping reliable machine learning systems into production. With 4+ years of experience, I build end-to-end pipelines, develop models, and craft API-based inference services that scale, perform, and stay maintainable. I have hands-on experience with LLMs, retrieval-based systems, and real-time prediction services, and I enjoy collaborating with cross-functional teams to integrate AI features with existing applications and business goals.

Available to hire

I’m Prakhar Mehta, an AI/ML engineer focused on shipping reliable machine learning systems into production. With 4+ years of experience, I build end-to-end pipelines, develop models, and craft API-based inference services that scale, perform, and stay maintainable.

I have hands-on experience with LLMs, retrieval-based systems, and real-time prediction services, and I enjoy collaborating with cross-functional teams to integrate AI features with existing applications and business goals.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

AI/ML Engineer at Cohere

January 1, 2025 - Present

Delivered end-to-end ML pipelines across three production use cases, improving system reliability through stronger validation, versioning, and structured deployment practices. Built retrieval-based systems using large language models and vector search, improving response accuracy by 38% and reducing information retrieval time by 45% across internal tools and workflows. Developed inference services using FastAPI and Docker supporting over 12K daily requests, maintaining consistent low latency and improving system stability through optimized request handling and efficient model loading. Automated model deployment workflows using Airflow and CI/CD pipelines, reducing release effort by 40% and enabling consistent, repeatable deployments across development, testing, and production environments. Improved model output consistency through structured prompt tuning and evaluation, reducing incorrect responses by 27% and strengthening reliability of AI-driven features used in daily operations. Bu

Associate AI/ML Engineer at CloudSphere IT Labs

July 1, 2020 - March 1, 2023

Transitioned into machine learning by contributing to predictive systems across multiple applications, reducing manual analysis effort by 30% and improving delivery of data driven features through structured modeling approaches. Built classification and regression models improving prediction accuracy by 22% on average, supporting use cases such as forecasting, segmentation, and performance analysis across client and internal projects. Developed feature engineering pipelines reducing data preparation time by 41%, enabling faster experimentation and improving model performance across structured datasets used in production environments. Created REST APIs for model inference enabling real time predictions, reducing response delays by 33% and supporting integration with backend services across internal and customer facing applications. Designed ETL pipelines improving data availability by 35% and ensuring consistent, reliable datasets for analytics and machine learning workflows across diff