Building AI from research.
Scaling it reliably.
I build AI organizations, platforms, and evaluation systems that turn ambitious research into reliable, production-scale products.
Background
My work sits at the intersection of AI research, production engineering, and organizational scale. I have spent more than a decade building capabilities where the path is not already obvious: bootstrapping low-resource language technology, taking multilingual NLU into production, building ML organizations and evaluation platforms, and creating systems that make modern AI more measurable and reliable.
At Sanas AI, I built and scaled the Data & ML Analytics organization to 40+ people and developed the operating model, data infrastructure, and evaluation systems supporting enterprise AI. At Amazon Alexa AI, I worked on multilingual production NLU, model quality, and Responsible NLU, contributing to programs including Alexa Teacher Model and the multilingual MASSIVE initiative.
Earlier, I bootstrapped low-resource NLP and Machine Translation programs for Bhojpuri, Maithili, and Magahi, coordinating ~60 project fellows across technical and linguistic workstreams. My work has also spanned speech corpora, STT/TTS resources, computational linguistics, sentiment and subjectivity analysis, AI bias, and model evaluation. I hold a PhD in Computational Linguistics/NLP from IIT (BHU).
My approach to AI leadership is simple: give research teams room to explore, but build the systems, evidence, and operating mechanisms needed to know rigorously what works and how to scale it. Outside work, I remain a student of Hindustani classical music; my interest in language and music comes from the same fascination with structure, variation, and meaning.
Professional Journey
Founder
Vaikhari AI • Bangalore
Building enterprise AI evaluation and reliability systems spanning benchmark design, reasoning, robustness, governance, and real-world task performance. Lead product strategy, research direction, platform architecture, and customer discovery.
Director, Head of ML Analytics
Sanas AI • Bangalore
Founded and scaled the Data & ML Analytics organization to 40+ people, building the team, operating model, large-scale analytics infrastructure, and Model Factory for automated data flows, experimentation, and model evaluation.
Senior Language Engineer / NLU Lead
Amazon (Alexa AI) • Bangalore
Led production NLU quality and multilingual language-engineering programs across global science and engineering teams. Bootstrapped and improved Indian English and Hindi NLU, drove data and evaluation programs, and contributed to Alexa Teacher Model and the 51-language MASSIVE initiative.
Senior Research Fellow & Teaching Assistant
IIT (BHU) • Varanasi
Led applied low-resource NLP and NLU research across data creation, modeling, experimentation, and evaluation. Taught NLP/DSP and mentored B.Tech students while researching sentiment and subjectivity analysis, multilingual NLP, and bias in language models.
Resource Person – Speech & Language Technology
LDC-IL / Central Institute of Indian Languages (CIIL) • Mysore
Worked across India’s scheduled languages on speech-corpus analytics and language-resource development, including aligned/segmented corpora and phonetic dictionaries for STT/TTS. Contributed to Hindi speech and text corpus development and advanced phonetics/phonology work.
Technical Lead & Project Coordinator
Project Varanasi, IIT (BHU) • Varanasi
Bootstrapped and led multiple low-resource NLP and Machine Translation programs for Bhojpuri, Maithili, and Magahi from the ground up—hiring, upskilling, and coordinating ~60 fellows while building corpora, annotation workflows, linguistic resources, and tooling for e-governance and underserved language communities.
Selected Research & Programs
Leadership & Technical Scope
My focus is not a single model family or research area. It is building the people, systems, and technical foundations required to turn AI research into repeatable product outcomes.
Organization Building
Applied AI
Natural Language Processing
AI Evaluation & Benchmarking
AI Reliability & Governance
Conversational AI
Generative AI
Bias in LLMs
Low-Resource Language NLP
Speech AI
ML Platforms
Responsible AI
Research → Production
Operating Scope
Zero-to-one teams, hiring, upskilling, org design, mentoring, and cross-functional leadership
Technical roadmaps, experimentation, evaluation, launch readiness, quality mechanisms, and model feedback loops
LLMs, NLU, speech, multilingual systems, model evaluation, data platforms, and large-scale ML pipelines
Let’s Connect
I welcome conversations around scaling AI organizations, applied AI strategy, model evaluation and reliability, multilingual and speech AI, and building zero-to-one AI platforms.