22 Nov
|
ChevalBleu Labs
|
Bengaluru
22 Nov
ChevalBleu Labs
Bengaluru
1. Role Overview ChevalBleu Labs is hiring a hands-on Senior AI/ML Engineer to build, deploy and maintain AI capabilities across our portfolio of products and platforms. This role covers generative AI, real-time conversational agents, retrieval-augmented systems, intelligent automation, and GPU-backed inference. You will take ownership of architecture, production reliability, model quality, latency, security and operating cost. 2.
Key Responsibilities
• Design and deploy production-grade generative AI, conversational agents, RAG applications and intelligent workflows for different products.
• Build real-time voice pipelines combining ASR, LLMs, TTS, VAD, turn-taking, interruption handling and barge-in.
• Integrate external AI services and self-hosted models through secure, reusable and provider-agnostic interfaces.
• Develop Python/FastAPI services, real-time WebSocket streams, REST APIs and event-driven backend integrations.
• Deploy and optimize GPU inference workloads; manage model serving, throughput, observability and cloud cost.
• Troubleshoot audio quality, noise suppression, echo cancellation, streaming crashes, timeouts and API failures.
• Improve performance through concurrency control, caching, batching, profiling and realistic load testing.
• Build evaluation pipelines, regression tests, security controls, monitoring, documentation and incident playbooks.
• Collaborate with product and engineering teams, review code, mentor developers and communicate technical risks. 3. Required Technical Expertise
• Backend engineering: Advanced Python, asynchronous programming,
FastAPI or equivalent, REST, WebSockets, Linux, Git and system design.
• LLMs and GenAI: Prompt engineering, tool/function calling, RAG, embeddings, vector search, agents, context management and evaluation.
• Voice AI: Streaming ASR/TTS, VAD, audio codecs, latency measurement, noise reduction, echo handling and production debugging.
• Infrastructure: Docker, AWS/GCP/Azure, GPU inference, CI/CD, logging, tracing, monitoring and secure secret management.
• Reliability: Retries, timeouts, rate limits, circuit breakers, graceful fallback, incident analysis and cost optimization. 4. Third-Party AI APIs and Integrations Practical experience integrating and evaluating multiple AI providers is expected. Familiarity with every vendor listed below is not mandatory.
• Voice and speech: ElevenLabs, Sarvam AI, Deepgram, OpenAI Whisper, Google Cloud Speech, Azure AI Speech or comparable providers.
• LLM APIs: OpenAI, Anthropic Claude, Google Gemini, Groq, OpenRouter or equivalent platforms.
• Telephony and real-time communications: Twilio, Plivo, Exotel, SIP, WebRTC, LiveKit or similar services.
• Orchestration and retrieval: LangGraph, LangChain, Hugging Face, FAISS, Qdrant, Pinecone or comparable tools.
• Integration engineering: Authentication, webhooks, streaming, vendor SLAs, rate limits, usage tracking and failover. 5. Self-Hosted AI and Advanced Capabilities
• Work with open-weight models such as Llama, Qwen and Mistral; deploy them using vLLM or similar serving stacks.
• Compare hosted APIs and self-hosted inference for quality, latency, data privacy, operations and total cost.
• Understand GPU/CUDA fundamentals; knowledge of quantization, batching, LoRA/QLoRA and model profiling is a plus.
• Experience with multilingual speech systems, including English and Hindi, is an advantage. 6. Qualifications
• 5+ years of relevant software or AI engineering experience, including ownership of deployed production systems.
• Strong hands-on coding, debugging and system design skills, with the ability to explain measurable outcomes.
• Bachelor's or master's degree in Computer Science, AI/ML or a related field, or equivalent practical experience.
• Experience leading complex technical deliveries and mentoring engineers is preferred. 7. What Success Looks Like
• Reliable AI systems across products, with measurable latency, error rates, model quality and operating costs.
• Fewer production incidents through testing, observability, clear fallbacks and end-to-end ownership.
• Modular architecture that allows AI providers and models to be changed without extensive rework.
• Repeatable deployments, practical technical documentation and predictable engineering delivery.
📌 Senior AI/ML Engineer (Bengaluru)
🏢 ChevalBleu Labs
📍 Bengaluru