Services
RAG pipelines, agentic workflows, Claude API integration, speech AI from desktop audio capture to speaker identity, document AI, and self-hosted or fine-tuned models on your own GPU infrastructure.
RAG pipelines, agentic workflows (LangChain/LangGraph), Claude API integration, speech AI (Whisper transcription, speaker diarization, native desktop system-audio capture on macOS and Windows, and speaker identity across Google Meet, Zoom and Teams), document AI and OCR, self-hosted and fine-tuned models (vLLM, LoRA/QLoRA on GPU infrastructure) and AI analytics.
Related work
AI & LLM Systems · Applications, Mobile & Backends
A phone and iPad assistant for a minister's Private Secretary that answers questions from government sources in Kannada or English, records meetings into minutes, triages email and turns it all into tasks the PS approves.
AI & LLM Systems · Solution Architecture
Five layers from ingestion to output, with domain adapters for healthcare, elections, FMCG and surveys, and forecasting and anomaly detection built in.
Healthcare AI · AI & LLM Systems · Solution Architecture
Seven automation areas across the revenue cycle, including ICD-CPT matching, document classification, OCR, insurance-portal automation and AI agents, delivered against a phased roadmap.
AI & LLM Systems · Solution Architecture
Telephony AI architecture for South Indian languages and English at 40,000 calls a day, covering the SIP stack, Indic speech recognition and synthesis, the LLM, the cost model and DPDP and TRAI compliance.
AI & LLM Systems
A deployment kit for vLLM on NVIDIA H200 GPUs, with LoRA and QLoRA fine-tuning and an OpenAI-compatible API.
Start with a conversation about the system you need, the constraints you have and where you are today.