Software Developer (Flutter Development with AI) in Vadodara for Night Shift

Job Description: Scope of Work: CuraTrail: Consumer-facing preventive health platform with HRV-based stress and readiness scoring (wearables and smart rings). Cross-platform mobile build using Flutter (Dart). Siyan Reflect: AI-powered CBT coaching app on a Swift / WebSocket / Azure voice pipeline, with two-layer memory injection, mood check-ins, session summaries, and pattern intelligence. Hybrid frontier-API + open-source-model architecture. AI Medical Receptionist: voice/chat agent integrating RingCentral (telephony), Tebra (EHR), and Availity (eligibility) for inbound calls, scheduling, intake, and insurance verification. Model Development: fine-tuning support (QLoRA on Llama 3.1 8B via Together AI), dataset curation, evaluation harnesses for clinical quality. Required Skills: Minimum 2 years of full-time, paid software development experience. Strong proficiency in Python (preferred), Node.js / TypeScript, or Go. Flutter (Dart) able to build, debug, and ship cross-platform iOS and Android apps. Production Flutter app published to App Store and/or Google Play (signing, release, review process). Front-end web experience with React or comparable framework. Hands-on REST API, webhook, and SDK integration. Git / GitHub workflows, code review, basic CI/CD. Working knowledge of one cloud provider (AWS, Azure, GCP, or DigitalOcean). Strong written and spoken English; able to communicate clearly during PST working hours. Strongly Preferred: iOS / Swift (relevant to Siyan Reflect); WebSocket real-time systems; voice/speech APIs (Azure Speech, ElevenLabs, Whisper). Fine-tuning workflows (QLoRA, LoRA, Together AI, Hugging Face); PostgreSQL / Supabase; Docker. Healthcare-API exposure (Tebra, RingCentral, Availity, FHIR); HIPAA / SOC 2 awareness. LLM API experience (Anthropic Claude, OpenAI, or equivalent) prompt design, tool/function calling, streaming. AI coding tools (Claude Code, Cursor, GitHub Copilot, Windsurf) used as a daily driver - able to drive prompts, review generated code critically, and override AI output when wrong. .