Subir Verma is an Applied Scientist II with a decade of experience turning research into production-grade AI across NLU, vector search, ranking, recommendation systems, and reinforcement learning. He has led and shipped ML solutions at Amazon, Freshworks, Tokopedia/GoTo, and Tata 1mg, blending pragmatic experimentation with robust MLOps (Airflow, Snowflake, Docker, AWS/Azure, FastAPI). Known for starting small—running careful experiments and iterating—he also mentors teams and aligns models to product needs, from conversational shopping assistants and probabilistic re-rankers to multimodal summarization and LLM-based evaluation. His background in signal processing and early work on audio watermarking and CNN-based fraud detection hints at strong foundations in both theory and applied engineering. Subir publishes and presents his work regularly and prefers collaborating closely with engineers and PMs to move scientific rigor into scalable features.
10 years of coding experience
8 years of employment as a software developer
Bachelor of Technology (BTech) Digital Signal processing (Specifically Audio Signal Processing) Circuit Designing & Engineering M, Bachelor of Technology (BTech) Digital Signal processing (Specifically Audio Signal Processing) Circuit Designing & Engineering M at The LNM Institute of Information Technology
High School Consumer EconomicsMathematicsPhysicsChemistryEnglish, High School Consumer EconomicsMathematicsPhysicsChemistryEnglish at Delhi Public School
This is MATLAB code for blind audio watermarking using Arnold transform and ECC
Contributions:9 commits, 2 PRs, 7 pushes in 4 years 2 months
audio-processingblindmatlabtransformwatermarking
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.