Open to opportunities

Anand Gupta

@anandgupta3

Message

I deploy safe LLMs and multilingual agents for healthcare, leading GenAI from vision to production.

Zimbabwe

Message

What I'm looking for

I’m looking for senior GenAI roles building reliable LLMs and multilingual agents in regulated or frontier applied-AI settings—where safety, evaluation/monitoring, and production ownership are core to the work.

I’ve spent 8 years building AI systems that actually ship in healthcare, where safety isn’t a checkbox. At Wysa, I rebuilt a rule-based chatbot into multilingual GenAI pipelines—LLM-agnostic architecture, RAG with guardrails, and multi-LLM failover—achieving >99% reliability and expanding from English-only to 6 languages (including three in three months). I cover the full stack of production NLP and LLM workflows, from Python and PyTorch to evaluation, alignment, and agentic systems, including LLM-as-judge pipelines that replaced 90% of manual QA.

Now as Head of AI at Wysa, I own AI strategy, roadmap, and responsible-AI guidelines for a global mental-health platform serving 3.5M+ users across high- and low-resource languages. I led a shift from traditional ML to LLM-powered conversational workflows, designing a multi-LLM-agnostic pipeline with p95 latency <5s, adaptive retries, circuit-breaker logic (Redis), and real-time deployment, while also building internal enablement tooling (an MCP plus a Streamlit platform) that cut AI engineering effort by 90%. My earlier work built safety-critical classifiers (85–95% accuracy) and a patented therapeutic alliance detection system, and my cybersecurity background shapes how I think about trust, failure modes, and what “safe enough” means in production.

Experience

Work history, roles, and key accomplishments

Current

Head of AI

Current

Wysa

Apr 2024 - Present (2 years 3 months)

Own the AI strategy, roadmap, and delivery for a global mental-health platform serving 3.5M+ users across multiple high and low resource languages. Lead an 8-member team spanning research, ML engineering, and MLOps.

Leadership & Strategy
- Define AI vision, OKRs, and responsible-AI guidelines for the team. Drove the strategic shift from traditional ML to LLM-powered workflows, modernising the e

AI Lead

Wysa

Feb 2024 - Apr 2024 (2 months)

Transitional role bridging individual contribution to AI leadership. Established evaluation frameworks, defined team processes, and laid the groundwork for the multi-LLM architecture.

- Designed and implemented Wysa's first LLM-as-judge evaluation pipeline, enabling automated QA for conversational models using persona-playing evaluator agents with chain-of-thought grading.
- Established developme

fastAPI Docker PyTorch Prompt Engineering Retrieval Augmented Generation (RAG)Transformers TensorFlow LLM As Judge MLOps Reliability (Retry And Failover) Logic

Data Scientist

Wysa

Sep 2021 - Feb 2024 (2 years 5 months)

Built ML systems powering Wysa's core conversational platform and drove multilingual global expansion.

ML Systems & Product Impact
- Delivered multilingual expansion (German, Spanish, Hindi, Marathi, Tamil, Bengali) end-to-end in under 3 months, enabling global rollout.
- Improved 25+ safety-critical classifiers (suicide risk, abuse, sentiment, crisis behaviours) to 85-95% accuracy, ensuring user

Monitoring Drift fastAPI Docker MLFlow PyTorch Prompt Engineering Retrieval Augmented Generation (RAG)Transformers TensorFlow ISO Aligned Audit Documentation GDPR Data Minimisation Model Monitoring Deep Learning

NLP Engineer

Wysa

Apr 2020 - Sep 2021 (1 year 5 months)

Worked on NLP systems for Wysa’s healthcare conversational platform, contributing to production-focused RAG workflows. Supported AI/ML engineering using FastAPI and TensorFlow-based components.

fastAPI Retrieval Augmented Generation (RAG)TensorFlow