Onkar Salvi
Senior Research Engineer
AI Research Engineer specializing in Natural Language Processing, Large Language Models, and AI Agents. Proven track record in designing cost-efficient, high-performance AI systems with measurable business impact.
About Me
I'm an AI Research Engineer with deep expertise in Natural Language Processing, Large Language Models, and AI Agents. Currently working as a Senior Research Engineer at Dataminr, where I focus on building scalable AI systems that deliver measurable business impact.
My work spans the entire AI lifecycle - from research and model training to scalable deployment and post-production optimization. I've published research in AAAI and have extensive experience in LLM-based automation, AI infrastructure, and model optimization.
I'm passionate about creating intelligent systems that can understand, reason, and interact naturally with humans while maintaining high performance and cost efficiency.
Experience
- Built and maintained company's in-house AI model serving framework with gRPC server and REST API proxy
- Achieved 70% reduction in throughput per dollar cost by migrating 14+ deep learning models to AWS Inferentia
- Designed Sales Enablement App resulting in 3× increase in email open rate and 60% boost in response rate
- Developed real-time threat detection system processing 600M+ daily messages with 23% F1 improvement
- Led migration of 40+ AI services from ECS to Kubernetes with substantial cost savings
- Research and developed various ML and DL models for NLP tasks including chatbots and reading comprehension
- Built Zero-shot Classification model achieving 98% accuracy on seen data and 90% on unseen data
- Used BERT and KMeans clustering on large call transcript datasets for semantic analysis
- Designed ranking models for multi-source news aggregation and relevance scoring
Skills & Technologies
🤖 Core ML and AI
⚡ Frameworks
☁️ MLOps & Infrastructure
Let's Connect
onkar.salvi7@gmail.com