Lead AI Engineer

About the Role

Venture AI builds AI and software systems for clients in the US and the Gulf. We are hiring an AI Engineering Lead to own the technical direction of our conversational AI products: the real time voice stack, the LLM and orchestration layer, telephony integrations, and the conversational analytics that show our clients what happened on every call.

You will lead a team of AI and backend engineers, set the architecture, and stay hands on. You will write code, review code, run client demos, and answer for delivery. You will also support our other client engagements, including a HIPAA compliant healthcare CRM, where your AI and data expertise will shape how automation and analytics are built.

This is a hands on leadership role: deep technical ownership alongside responsibility for your team and for client outcomes.

What You’ll Do

Voice AI and real time agents

• Own the architecture of our real time voice agent stack, built on LiveKit (rooms, tracks, agents framework, SIP) and streaming STT, LLM and TTS pipelines.
• Design for conversational quality: end to end latency budgets, voice activity detection, turn taking and interruption (barge in) handling, endpointing, and graceful recovery from silence, noise and poor audio.
• Evaluate, integrate and swap STT and TTS providers (for example Deepgram, AssemblyAI, Whisper, ElevenLabs, Cartesia or similar) on measured accuracy, latency, cost and language coverage, including Arabic and accented English for Gulf and US clients.
• Build and tune LLM driven dialogue: prompt and tool design, function calling, structured outputs, guardrails, fallbacks and human handoff.

Telephony orchestration

• Architect inbound and outbound calling across SIP trunks and CPaaS providers (such as Twilio, Telnyx or similar), including call routing, transfers, call event webhooks, retries and reconciliation.
• Integrate agents with client systems (CRMs, scheduling, ticketing and CTI platforms) so calls trigger real actions, not just conversation.
• Run voice workloads reliably at scale: concurrency, autoscaling, failover and cost control per call minute.

Conversational analytics and RAG

• Build the analytics layer on top of every conversation: transcription post processing, summarization, intent and topic extraction, sentiment, compliance and QA scoring, and outcome tracking.
• Design RAG pipelines (chunking, embeddings, hybrid retrieval, reranking, grounding and citation) that give agents accurate, client specific knowledge in real time and power post call insight.
• Build agentic workflows with LangGraph, LangChain or equivalent frameworks, with clear state, tools, and testable behaviour.
• Stand up evaluation and observability for LLM and voice systems: offline test sets, LLM as judge scoring, regression suites, tracing and monitoring (for example Langfuse, LangSmith or similar).

Leadership and delivery

• Lead and mentor AI, backend and frontend engineers: plan the week, assign work, review every pull request, and raise the team’s standard.
• Run milestone demos with clients, hold scope against written acceptance criteria, and track risks and dependencies before they block delivery.
• Estimate and scope new phases and change requests on fixed price, milestone based engagements.
• Deploy and operate on AWS with isolated environments and CI/CD, with security, audit logging and data privacy controls built in.

Supporting other Venture AI projects

• Contribute to the healthcare CRM engagement (FastAPI, PostgreSQL, Airflow, HIPAA controls) and to Python and LLM API pipelines that process deidentified clinical data, applying the same engineering standards.

Requirements

• 6+ years of professional software engineering experience, including 2+ years leading a team and 2+ years building production AI or LLM based systems.
• Bachelor’s or Master’s degree in Computer Science, Software Engineering or a closely related discipline.
• Hands on production experience with voice AI: real time STT, LLM and TTS pipelines, and the tradeoffs between latency, quality and cost.
• Strong experience with LiveKit (agents framework, rooms, SIP) or comparable real time communication stacks such as WebRTC.
• Practical telephony knowledge: SIP, PSTN, call flows, and integration with a CPaaS or carrier such as Twilio or Telnyx.
• Strong LLM application skills: prompt and tool design, function calling, structured outputs, and evaluation.
• Experience building RAG systems and agentic workflows with LangGraph, LangChain or similar, including vector search (for example pgvector or a managed vector store).
• Strong Python backend skills: FastAPI or an equivalent framework, async and streaming APIs, WebSockets, Pydantic, and automated testing.
• Solid PostgreSQL and data pipeline fundamentals; production experience on AWS with Docker and CI/CD (GitHub Actions or similar).
• Sound grasp of application security: authentication, role based authorization, audit logging and secrets management.
• A record of shipping client projects on time, and clear written and spoken English for direct work with US and Gulf clients.
• Availability to overlap with US Central Time for standups and demos.

Bonus Skills (Nice to Have)

• Experience with conversation analytics at scale: call QA, speaker diarization, sentiment, topic clustering and dashboards.
• Arabic or multilingual voice agent experience, including dialect handling.
• Fine tuning or distillation of models for speech or dialogue, and experience with open source and self hosted models.
• Experience with HIPAA regulated systems or other PHI/PII sensitive environments, or healthcare data (EHR, ICD-10, Medicare programs).
• Experience with CTI or contact centre platforms.
• Enough frontend fluency to review a junior engineer’s work and make UI architecture decisions.

Why Join Venture AI

• Lead the technical direction of a production voice AI product used on real customer calls.
• Work with the full modern voice and LLM stack, from real time audio to analytics, rather than a single narrow slice.
• Work directly with US and Gulf clients on systems their teams rely on every day.
• Real ownership of architecture, team and delivery.
• Competitive compensation with performance bonus, learning allowance and medical allowance.