KG

Principal SWE, Applied AI

Krishnam Gupta

I build AI that ships to millions in healthcare - from diagnostic pipelines behind 1.3M+ tests across 12+ countries to LLM systems guiding 50,000+ users through care.

Krishnam Gupta

About

Principal Software Engineer, Applied AI at Audere. 9+ years building production AI across computer vision, LLMs, and robotics. Published researcher (7 papers, 1 patent). Peer reviewer for Nature Communications, IEEE, and ACM venues. MS + BTech from IIIT Hyderabad.

Computer VisionMedical AILLM SystemsLLM EvaluationRoboticsAI for Global Health

Experience

Audere

2022 - Present

Principal SWE, Applied AI - Seattle, WA (Remote from SF)

  • Architected CV pipeline behind 1.3M+ rapid diagnostic tests (malaria, HIV, COVID) across 15+ countries
  • Led rule-based to deep learning migration: sensitivity 82% to 94%, specificity >99%
  • Designed structured extraction from 30,000+ WhatsApp conversations (120+ clinical fields)
  • Built LLM evaluation framework benchmarking 11+ models, cut costs 50-63%
  • Designed multi-provider LLM fallback with open-weight models at 80%+ lower cost, with production guardrails including content moderation and clinician escalation
  • Co-authored 3 peer-reviewed papers on diagnostic AI while shipping production systems

Mfine

2019 - 2022

Lead Data Scientist - Bangalore, India

  • AI differential diagnosis system serving 100K+ consultations/month
  • Skin disease recognition: SOTA +5% (ICPR 2020 paper)
  • Led team of 4 data scientists across CV and NLP
  • Built smartphone-based PPG system for pulse rate and oxygen saturation - remote vitals without specialized hardware (patent pending)

Microsoft Bing

2017 - 2019

Research Engineer - Hyderabad, India

  • Search relevance and ranking models serving billions of daily queries
  • Developed ML features for query-document matching and click-through rate prediction at web scale

Publications & Patents

Healthcare AI

Robotics & Vision

Patents

Differential Diagnosis System

US20220084677A1 - Granted

Awards & Peer Review

  • -Stanford T.H.I.N.K 2026 - Poster presentations on LLM evaluation and clinical data extraction
  • -Finalist, 12th Aegis Graham Bell Awards - Innovation in Diagnostics
  • -Dean's Merit List - All 4 undergraduate years, IIIT Hyderabad
  • -Reviewer, Nature Communications
  • -Reviewer, CADL Workshop at ICPR 2020
  • -Reviewer, SASIGD 2026
  • -Invited Reviewer, AIiH 2026 (Imperial College London)

Projects

SitePulse

AI-powered 3D site assessment for humanitarian clinics. Three.js + Gaussian Splats + Claude.

Next.js, Three.js, Claude API

VLA Edge

Toolkit to profile, optimize, and deploy Vision-Language-Action models to edge hardware like Jetson Orin Nano.

Python, PyTorch, TensorRT, llama.cpp

Physical AI

Teaching robots to pour 50ml via imitation learning. LeRobot + ACT policy.

Python, LeRobot, ACT

Impact

1.3M+

diagnostic tests powered

12+

countries deployed

50,000+

users served

30,000+

LLM conversations

7

publications, 65 citations

Let's Connect

Open to conversations about applied AI, healthcare technology, research collaborations, and frontier AI roles.