Open to opportunities

Raghav Mehta

@raghavmehta

Message

Senior data scientist specializing in computer vision, scalable multimodal search, and model optimization.

India

Message

What I'm looking for

I seek roles building production-grade computer vision and multimodal systems, with opportunities to lead research-to-product delivery, mentor engineers, and drive measurable enterprise impact.

I am a senior data scientist focused on computer vision research, scalable multimodal systems, and model optimization. I lead product-focused research efforts and have scaled teams while collaborating directly with executive leadership.

At my current company I developed efficient explainability techniques, billion-scale multimodal visual search used by enterprise customers, and auto-labeling libraries that cut manual labeling costs by 80%. I also architected cloud training infrastructure and optimized models for real-time deployment.

My research includes synthetic data generation, domain adaptation, and few-shot visual anomaly detection with state-of-the-art results; I have deployed solutions for customers including aerospace, medical device, and rail industries. I co-authored peer-reviewed publications and have multiple patent filings in dataset curation and VLM-based annotation.

I seek to drive impactful, production-ready AI systems that balance research rigor with operational efficiency, mentor engineering talent, and deliver measurable ROI for enterprise customers.

Experience

Work history, roles, and key accomplishments

Current

Senior Data Scientist

Current

Akridata

Jan 2024 - Present (1 year 9 months)

Led computer vision research and core product development, scaling the team to 12 and delivering billion-scale multimodal visual search and auto-labeling solutions that reduced labeling costs by ~80% and cut saliency compute time by 75%.

Computer Vision Deep Learning PyTorch Onnx Synthetic Data Multimodal Retrieval

Data Scientist

Akridata

Jun 2021 - Jan 2024 (2 years 7 months)

Contributed to core CV product features including few-shot visual anomaly detection, VLM-based auto-labeling, and inference optimizations (INT8, pruning, distillation) that enabled real-time deployments and ~30% inference cost reduction.

Few Shot Learning Anomaly Detection Pruning Knowledge Distillation TensorRT Onnx Deployment

Data Scientist

The Solar Labs

Aug 2019 - Jan 2021 (1 year 5 months)

Trained and deployed DeepRoof, a two-stage segmentation ensemble for LiDAR and SfM height maps achieving 95% building footprint coverage and built end-to-end raster/vector pipelines and ML infrastructure for geospatial analysis.

Remote Sensing Segmentation LiDAR Data Pipelines Image Processing Geospatial analysis

Research Assistant

Jaypee Institute of Information Technology

Jan 2018 - Jan 2020 (2 years)

Co-authored three peer-reviewed papers and developed an ensemble deep learning model that improved web navigation prediction accuracy by 7.67% on the Wikispeedia dataset.

Research Deep Learning Sequence Modeling Data Visualization Experimentation Paper Writing

Data Science Intern

Instalocate

May 2018 - Sep 2018 (4 months)

Built models combining ARIMA and LSTMs to predict flight delay causality and ripple effects, and developed a Django+Plotly analytics platform that increased client conversion and a Twitter bot that drove rapid engagement.

Python Django Plotly Data Engineering Experimental Analysis Product Analytics

Education

Degrees, certifications, and relevant coursework

Jaypee Institute of Information Technology

Bachelor of Technology, Computer Science and Engineering

Activities and societies: Guest lecturer and workshop instructor on deep learning; involved in research projects and publications on web navigation prediction and deep learning-based sequence modeling.

Completed Bachelor of Technology in Computer Science and Engineering, focusing on deep learning, model optimization, and applied machine learning techniques.