AI/ML Engineer (Voice AI and Agentic Systems)

About Us

GTechSources is the Lahore-based technology arm of a US company, with a team of 60+ engineers, QA specialists and product people. We build and run software products for the US veterinary industry, and AI now sits at the centre of what we ship.

Our AI work is in production, not on slides. We run an AI receptionist that answers real phone calls for veterinary clinics, books appointments and triages symptoms; an AI system that turns consultations into clinical SOAP notes; an AI call agent backed by a large veterinary knowledge base; and upcoming work on product training and demand forecasting. We are looking for an engineer who wants to own these systems and make them faster, smarter and more reliable.

The Role

You will design, build and maintain AI features end to end, with a strong focus on real-time voice agents and agentic workflows. You will work closely with our web, mobile and QA teams, integrating your services with our Laravel, Next.js and Node.js platforms. This role suits someone who enjoys solving hard problems, can track down a bug across a live call pipeline, and cares about how the product behaves for a real caller on the other end of the line.

Key Responsibilities

• Build and improve real-time voice AI agents: speech-to-text, LLM reasoning, text-to-speech, and telephony, with low latency and natural conversation flow.
• Handle the hard parts of live calls: turn detection, interruptions (barge-in), silence handling, call transfers and graceful failure.
• Design agentic and multi-agent workflows using tool/function calling, structured outputs and conversation state management.
• Build and tune RAG pipelines over domain knowledge bases, including chunking, embeddings, retrieval quality and grounding.
• Work on transcription and clinical note generation, improving accuracy, structure and consistency of outputs.
• Write prompts and evaluation sets, measure quality, and reduce hallucinations and failure cases in production.
• Build clean Python services and APIs (FastAPI or similar) that integrate with our PHP/Laravel and Next.js applications.
• Debug production issues across the full pipeline using logs, call recordings and traces, and fix root causes rather than symptoms.
• Monitor cost, latency and reliability of AI services, and choose the right model or provider for each task.
• Work with our SQA team to define test cases for AI behaviour, and support releases through testing and deployment.
• Keep up with new models and tools, and propose practical improvements to our products.

Required Skills and Experience

• 3+ years of hands-on experience in AI/ML engineering, with at least one LLM-based product shipped to real users.
• Strong Python skills, including async programming (asyncio), REST APIs and WebSockets.
• Practical experience with LLM APIs such as OpenAI, Anthropic Claude or Google Gemini, including function calling and structured outputs.
• Experience building voice or conversational AI, using STT and TTS services (for example Deepgram, Whisper, ElevenLabs or Google).
• Solid understanding of RAG: embeddings, vector databases (pgvector, Pinecone, Qdrant or similar) and retrieval evaluation.
• Strong debugging and problem-solving ability. You can read unfamiliar code, reproduce issues and find the real cause.
• Working knowledge of SQL NoSQL databases (PostgreSQL and/or MySQL, MongoDB).
• Comfortable with Git, Docker and deploying services to cloud platforms.
• Good written and spoken English, as you will work with US-facing products and stakeholders.

Nice to Have

• Telephony experience: Twilio, SIP, WebRTC, or voice frameworks such as Pipecat, LiveKit or Vapi.
• Experience with agent frameworks such as LangGraph, LangChain, CrewAI or the OpenAI Agents SDK.
• Knowledge of VAD and turn-detection models, streaming audio and latency optimisation.
• Experience with classical ML or time-series forecasting (scikit-learn, Prophet, XGBoost).
• Exposure to healthcare or veterinary data and related privacy practices.
• Familiarity with PHP/Laravel, Node.js or Next.js codebases.
• Experience on DigitalOcean, AWS or GCP.

What We Offer

• Market-competitive salary.
• Weekends off (Saturday and Sunday).
• 22 paid holidays.
• Medical reimbursement.
• Referral bonus.
• Overtime payment.
• Work on live AI products used by businesses in the US market.
• Real ownership of systems, with direct access to leadership and quick decision-making.
• Competitive salary based on skills and experience.
• A supportive team of 60+ professionals across development, QA and product.
• Exposure to the latest models, tools and AI infrastructure.
• Clear growth path into senior and lead AI roles.

Hiring Process

• Step 1: CV screening.
• Step 2: Technical interview, a discussion around your past projects and AI/ML fundamentals.
• Step 3: One-hour on-site hands-on assessment covering a practical task, logic building and bug hunting.
• Step 4: Final discussion and offer.

Working Hours:

• 12:00 PM – 9:00 PM

How to Apply

Apply through Indeed with your updated CV. Please include links to your GitHub, portfolio or any AI projects you have built, and a short note on the most challenging AI problem you have solved. Candidates who can show working voice or agentic AI projects will be given priority.

Pay: Rs50,000.00 - Rs100,000.00 per month

Ability to commute/relocate:

• Lahore: Reliably commute or planning to relocate before starting work (Preferred)

Application Question(s):

• What is your current salary?
• What is your salary expectation?
• Please note that this is an onsite position, and our standard operating hours are from 12 PM to 9 PM, Monday to Friday. Are you available and willing to work during these hours?

Experience:

• AI: 3 years (Required)
• ML: 3 years (Required)
• Voice AI: 3 years (Required)
• Agentic Systems: 3 years (Required)
• FastAPI: 3 years (Required)
• Python:: 3 years (Preferred)

Location:

• Lahore (Preferred)

Work Location: In person