Case Studies
How GenAI Protos turns AI models into governed, API-ready systems that teams can operate.
Private Multilingual Transcription API Deployed on DGX Spark
GenAI Protos deployed Whisper-large-v3 as a private multilingual speech-to-text API on NVIDIA DGX Spark. The service uses vLLM, exposes an OpenAI-compatible audio transcription endpoint through the shared gateway, and supports batch transcription workflows across 99 languages without forcing teams to call a public transcription API.
Private 3D Generation API Deployed on DGX Spark
GenAI Protos deployed TRELLIS as a private 3D generation API for image-to-3D and text-to-3D workflows. The service returns textured 3D assets, supports synchronous or asynchronous job handling, and uses on-demand model loading with idle unload to manage resource use.
Private Multimodal LLM API Deployed on DGX Spark
GenAI Protos deployed Gemma-4-26B-A4B-it as a private LLM service on NVIDIA DGX Spark. The build used vLLM with FP8 runtime quantization, right-sized GPU allocation, OpenAI-compatible chat access through the shared gateway, and warm-up steps to improve real request readiness.
Private Long-Form Audio Transcription API Deployed on DGX Spark
GenAI Protos deployed a private long-form transcription service on NVIDIA DGX Spark using a Qwen2.5-7B-based VibeVoice-ASR model served through vLLM and exposed through a governed LiteLLM gateway. The service accepts audio as an OpenAI-compatible chat-completions request, streams structured transcript output, and supports long recordings, speaker labels, timestamps, multilingual audio, and inference-time hotword biasing.
Private AI Video Generation API Deployed on DGX Spark
GenAI Protos deployed a private video generation service on NVIDIA DGX Spark, exposed it through a LiteLLM gateway, and documented text-to-video, image-to-video, async job, preload, unload, and health routes. The implementation supports MP4 generation through custom pass-through endpoints while using on-demand loading and idle auto-unload to manage GPU memory.
Private Multi-Mode Text-to-Speech API Deployed on DGX Spark
GenAI Protos deployed a private multi-mode text-to-speech system that supports preset voices, instruction-based voice design, and reference-audio voice cloning. The system uses a FastAPI orchestrator in front of multiple vLLM-Omni backends and exposes OpenAI-compatible speech access through the GenAI Protos gateway.
Private Coding and Reasoning LLM API Deployed on DGX Spark
GenAI Protos deployed a private coding and reasoning LLM service on NVIDIA DGX Spark using vLLM and FP8 quantization. The build exposes OpenAI-compatible chat completions through a governed gateway, supports streaming, tool-use patterns, and code-review workflows, and uses warm-up steps to reduce first-request latency after model start.
Private Multilingual Voice AI API Deployed on DGX Spark
GenAI Protos deployed OmniVoice on NVIDIA DGX Spark as a private multilingual voice AI service. The build exposed OpenAI-compatible speech generation, custom voice cloning and voice design workflows, streaming synthesis, language and voice-attribute discovery, health checks, gateway authentication, and documented runtime behavior for production-style evaluation.
AI-Powered Retail Personalisation Platform
GenAI Protos designed and delivered an AI-powered Retail Personalisation and Fulfilment Platform for a global-scale retail enterprise. The platform unifies clickstream, search, purchase history, visual signals and cart events into a real-time customer intent layer. It combines deep learning, regression models, multimodal recommendations and agentic fulfilment orchestration to improve discovery, checkout and post-purchase engagement, with documented outcomes including a 35% incremental revenue uplift and 20% customer loyalty improvement.
Real-Time Social Media Sentiment Intelligence Platform
GenAI Protos engineered a production-grade Agentic AI Sentiment Intelligence Platform for a financial services and brand intelligence use case. The platform processes 2.3M+ daily social signals, uses a multi-agent LangGraph workflow, grounds sentiment reasoning through Advanced RAG, and delivers 94% sentiment accuracy, sub-800ms p95 signal-to-insight latency, 67% analyst time savings and 3.1x ROI at six months.
AI-Powered Marketing Intelligence Platform
GenAI Protos built a multi-agent AI marketing intelligence platform that reduced content creation time by 50%, improved campaign ROI by 20%, accelerated audience segmentation by 3x and increased engagement by 40%. The platform uses specialised agents for content generation, consumer trend analysis and personalised advertising, coordinated by an orchestrator agent and grounded by RAG, CRM data and analytics signals.
AI-Powered IT Operations Intelligence
GenAI Protos designed and deployed an AI Operations platform that unifies infrastructure telemetry, logs, metrics, and event streams into an intelligence layer for proactive monitoring, anomaly detection, predictive maintenance, intelligent ticketing, and L1 support automation. The system uses Azure AI services, Azure ML, Azure Event Hubs, Logic Apps, ITSM integrations, and an AI support bot to reduce IT support cost, improve uptime, accelerate ticket resolution, and deflect repetitive L1 incidents.
AI-Powered Contract Intelligence: Transforming Legal Document Review with GenAI
GenAI Protos designed and deployed an AI-powered contract intelligence platform for a legal services organisation handling thousands of contracts annually. The platform combines Large Language Models, Retrieval-Augmented Generation, multimodal document parsing, semantic clause extraction, agentic risk intelligence, and reviewer validation to reduce contract review time by 40%, improve risk detection, reduce billable overhead, and provide clause-level review visibility for legal teams.
AI-Powered Radiology Intelligence Platform
GenAI Protos designed and deployed an AI radiology intelligence platform for a leading healthcare organization. The platform fuses imaging data, patient history, and lab results to support radiologist review, surface ranked differential diagnoses, generate structured reports, and automate adjacent workflow steps. The documented impact includes faster diagnosis, reduced unnecessary procedures, less documentation time, and higher radiology capacity.
AI-Powered Adaptive Learning Platform
GenAI Protos designed and delivered an AI-powered Adaptive Learning Platform for a leading online learning enterprise. The production system continuously analyses learner performance, engagement, and behavioural signals to personalise content modality, difficulty, and sequence in real time. The platform combines deep learning learner modelling, an adaptive difficulty engine, multimodal content delivery, AI-assisted content generation, and feedback optimisation - producing measurable gains in enrolment, completion, engagement, learner satisfaction, and content efficiency.
Private Voice Conversion API Deployed on DGX Spark
GenAI Protos deployed a private RVC voice conversion service that applies a trained reference voice model to uploaded audio or generated speech. The service wraps the RVC v2 inference pipeline in FastAPI, exposes custom gateway routes, and includes preload, unload, health, and model listing controls for operational use.
Private Full-Duplex Voice Agent Runtime Deployed on DGX Spark
GenAI Protos deployed PersonaPlex-7B as a private full-duplex speech-to-speech runtime on NVIDIA DGX Spark. The service supports real-time bidirectional voice interactions, persona control through system prompts, and 18 voice variants, using direct WebSocket access rather than a standard OpenAI HTTP endpoint.
Private AI Music Generation API Deployed on DGX Spark
GenAI Protos deployed a private text-to-music generation service on NVIDIA DGX Spark. The implementation wraps HeartMuLa-oss-3B and the HeartCodec decoder in a FastAPI service, exposes the capability through a LiteLLM gateway, supports synchronous and asynchronous MP3 generation, and includes health, preload, unload, lazy-load, and auto-unload controls for practical operation.