Looking for a job

Sarvesh Upasani

@sarveshupasani

AI/ML Engineer focused on scalable machine learning systems, real-time data pipelines, and production-ready model deployment.

India

Message

What I'm looking for

I’m looking to build scalable, high-performance AI systems in production—real-time data pipelines, strong model deployment/MLOps, and large-scale inference. I want to optimize latency and relevance with measurable impact.

I’m an AI/ML Engineer specializing in scalable machine learning systems, real-time data pipelines, and distributed architectures. I’ve built AI applications with a strong focus on model deployment, system optimization, and large-scale data processing—creating measurable impact through high-throughput systems and document intelligence solutions.

At Hindustan Aeronautics Limited (HAL), I developed a RAG-based intelligent search system over 50,000+ technical documents, reducing manual search time by 70% and improving retrieval relevance by 35%. I also built ML models for satellite-based crop classification on 100K+ samples at MITCON, improving accuracy by 18% and reducing data preparation time by 40%, and I’ve delivered systems like a real-time anomaly detection pipeline (1M+ events/day, 92% precision) and scalable semantic search with vector-based retrieval.

Experience

Work history, roles, and key accomplishments

AI ML Intern

Hindustan Aeronautics Limited (Hal)

Jan 2025 - Jan 2026 (1 year)

Developed a RAG-based intelligent search system over 50,000+ technical documents, reducing manual search time by 70% and improving retrieval relevance by 35%. Built semantic search pipelines for context-aware document querying and designed a scalable enterprise-ready document intelligence system.

RAG Semantic Search NLP Machine Learning Pipelines Data Processing Python Apache Spark

AI ML Intern

Mitcon Consultancy & Engineering Services Ltd

Jan 2026 - Present (5 months)

Built machine learning models for satellite-based crop classification on 100K+ data samples, improving accuracy by 18% through feature engineering and hyperparameter tuning. Automated preprocessing pipelines, reducing data preparation time by 40%.

Hyperparameter Tuning Model Training data preprocessing Classification Python SQL