Bhimesh Chauhan

Architected and built a multi-tenant SaaS platform with a production RAG chatbot, owning the system from retrieval architecture through backend delivery and reliable releases.

• Led Ask Kindred from SDK integration through production, serving approximately 200 daily active users across three customer companies with grounded retrieval, citations, conversation history and feedback.
• Built RareDaysAI, the reusable multi-tenant RAG platform behind it, shipped as an npm SDK and CLI with config-as-code tenants, staging-then-approval production deploys, per-tenant capacity caps and per-client cost attribution. About 90% of the repository's commits are mine.
• Runs it at under 5 seconds to first token and 99.99% availability for about $15 per client per month.
• Evaluates retrieval and refusal behaviour against fixed question sets: a 55-query set for Ask Kindred (distress queries routed correctly 7/10 to 10/10, all out-of-scope queries refused with no grounding violations).
• Led engineering for the Jefferson Fisher launch platform: about $700K USD in total sales in under 48 hours.
• Designed the multi-tenant SaaS architecture, including backend infrastructure and role-aware access controls.
• Delivered 10+ CI improvements that strengthened release quality and platform reliability.
• Mentored a junior engineer and created reusable AI capabilities for other Rare Days projects.

Founded and led all engineering for an ambient clinical AI platform that structures doctor-patient conversations into medical documentation.

• Architected LLM and RAG workflows for real-time transcription and structured note generation, reducing clinician documentation time by about 80%.
• Set the technical roadmap from MVP through production, balancing clinical workflow quality, security and reliability.
• Worked with clinicians to validate outputs and guide the evolution of the AI system.
• Led a team of 4-5 engineers and designers.

Led AI data-platform engineering for biomedical research workflows.

• Built AI-driven pipelines processing 100K+ research artifacts into structured datasets.
• Reduced manual data curation by about 60% by standardizing AI-assisted extraction and validation.
• Designed RAG workflows for literature synthesis and discovery; developed Neo4j and PostgreSQL query paths.
• Established data reliability, lineage and auditability practices across engineering, science and ML teams.

Built and operated distributed software for smart food kiosks.

• Improved customer engagement by 20% and satisfaction by 15%, and reduced downtime by 30%.
• Implemented Modbus RS485 integrations in Python and Node.js connecting cloud services to embedded hardware.
• Improved database query performance by about 25% and reduced bug reports by about 40%.
• Delivered production payment workflows and supported payment-certification work.

Built personalization and analytics capabilities for a global language-learning marketplace serving 80K+ learners.

• Led a cross-functional team delivering student progress-tracking and feedback for 80K learners.
• Built NLP-driven personalization that increased engagement by 20% and satisfaction by 15%.
• Ran tests and product analysis across markets contributing to a sustained 5% increase in conversion.

Built enterprise integrations connecting AutoML workflows to customer data and analytics tools.

• Integrated DataRobot with Excel, Snowflake, Tableau and Alteryx, increasing partner visibility by 70%.
• Built custom Jenkins and Kibana CI/CD tooling for repeatable ML-backed releases.

Built SaaS applications for financial and legal teams.

• Built Node.js REST APIs for high-throughput client workflows adhering to OWASP and HIPAA requirements.
• Developed two GraphQL-powered MeteorJS applications.

Built software for connected aeroponics systems.

• Built an Android application to monitor and control aeroponics firmware through MQTT.
• Developed PyTorch and TensorFlow models for resource prediction, increasing crop yield by 20%.