AI/ML Lead

Amex Hub; Islamabad; On-site; PKR 6,000,000 - 10,000,000; Full time

Job summary

Amex Hub leads innovative AI/ML initiatives across production systems in Islamabad. The role demands end-to-end ownership from identifying opportunities to deployment and scaling, with a strong emphasis on designing LLM/NLP systems and robust MLOps.

You will mentor engineers, set architectural standards, and drive cost-efficient, low-latency solutions while collaborating with product and leadership to deliver measurable business value.

Qualifications

• 6+ years building and shipping AI systems in industry.
• Track record of 3–5 substantial AI/ML production projects.
• Deep Python experience with PyTorch or TensorFlow, plus SciKit-Learn, NumPy and Pandas.
• Strong knowledge of LLM/NLP systems, prompts, tool-calling and model selection.
• Design and evaluate RAG systems with embeddings and vector stores.
• Experience with MLOps, CI/CD, monitoring, and Docker; cloud GPU infra.
• SQL and Linux/Bash proficiency; strong product and business sense.
• Experience mentoring engineers and leading AI/ML teams.

Responsibilities

• Own AI/ML initiatives end-to-end from ideation to production.
• Decide on model selection, architecture, and deployment trade-offs.
• Lead design of LLM/NLP systems, RAG pipelines and vector stores.
• Define evaluation strategies, benchmarks, and guardrails.
• Drive fine-tuning vs prompting decisions and lead tuning efforts.
• Own deployment/inference architecture optimizing latency and cost.
• Set standards for code quality, MLOps, monitoring, and experiments.
• Mentor engineers and review designs and code.

Skills

Leadership

Production AI/ML

Python

ML fundamentals

LLM systems

MLOps

AWS/GCP/Azure

SQL

Linux/Bash

Product & business sense

Mentoring

Communication

Tools

PyTorch

TensorFlow

Scikit-learn

NumPy

Pandas

LLM tooling

RAG pipelines

Pinecone/FAISS/Weaviate

Docker

Kubernetes

MLflow

Job description

Lead and ship production-ready AI/ML solutions, from identifying and validating opportunities through architecture, deployment, and scale.

Engineering Bahria Town Islamabad Full-time

About the role

Responsibilities

• Own AI/ML initiatives end-to-end, from opportunity identification and validation through design, proof of concept, production deployment, and scaling
• Make independent technical decisions on model selection, architecture, infrastructure, and trade-offs; flag risks early, own mistakes, and drive fixes without waiting to be asked
• Lead the design and implementation of LLM/NLP systems, including RAG pipelines, embeddings, retrieval, and reranking
• Define the evaluation strategy: metrics, benchmarks, and guardrails that reflect real product quality, not just academic accuracy
• Decide when fine-tuning is the right call versus prompting, RAG, or off-the-shelf models, and lead fine-tuning efforts when it is
• Own deployment and inference architecture, optimizing for latency, cost, and reliability
• Set team standards for code quality, experimentation rigor, MLOps, and monitoring
• Mentor engineers and lead technically: review designs and code, and raise the team's ML engineering practices
• Make build-vs-buy decisions on models, tools, and infrastructure (open-source vs. proprietary LLMs, managed vs. self-hosted inference)
• Work with product and leadership to turn business problems into AI solutions and set realistic expectations about what AI can and cannot do
• Identify AI/ML opportunities based on market needs, customer problems, competitive landscape, and emerging trends.
• Validate AI product ideas through prototypes, user feedback, experiments, and measurable product/business outcomes before scaling.
• Assess the commercial viability of AI solutions, including user value, adoption potential, operating / model costs, and scalability.

Required Skills & Experience

• 6+ years of experience, mostly building and shipping AI systems in industry rather than in purely academic or research settings
• A track record of 3–5 substantial AI/ML projects in real production use (not all need to be LLM-based)
• Production-level Python with deep hands-on experience in PyTorch and/or TensorFlow, plus Scikit-learn, NumPy, and Pandas
• Deep grasp of machine learning and deep learning fundamentals
• Strong experience with LLM/NLP systems: open-source and API-based models, prompt engineering, structured outputs, function/tool calling, and model selection
• Practical experience designing RAG systems: embeddings, semantic/hybrid search, retrieval and reranking, and vector databases (Pinecone, Weaviate, FAISS, pgvector, Milvus, or similar)
• Hands-on fine-tuning experience, including parameter-efficient methods such as LoRA/PEFT
• Proven ability to design evaluation frameworks, benchmarks, and guardrails, and to catch failure modes before and after deployment
• Experience with MLOps and production practices: model serving, monitoring, experiment tracking, CI/CD, and Docker
• Hands-on experience with AWS, GCP, or Azure, including GPU training/inference infrastructure, scalability, reliability, and cost optimization
• Proven ability to design AI system architecture from scratch and take projects from proof of concept to production
• Working knowledge of SQL and Linux/Bash
• Strong product and commercial awareness, with the ability to assess whether AI/ML solutions address meaningful market or customer needs and deliver measurable business value.
• Experience leading or mentoring engineers
• Strong communication skills: able to explain technical trade-offs clearly to engineers and non-technical stakeholders
• A strong sense of ownership: treats production systems as theirs to protect, follows through on commitments, and doesn't leave problems for others to catch

Preferred Skills & Tools

• Designing agentic workflows, tool-using AI systems, or multi-agent architectures (highly valued)
• Computer vision systems, including enough depth to review CV work done by the team
• MLOps platforms: MLflow, Kubeflow, Weights & Biases, or equivalent
• Model serving frameworks: vLLM, Triton Inference Server, or equivalent
• Infrastructure and data pipelines: Kubernetes, Spark, Airflow, or equivalent
• LLM tooling: Hugging Face Transformers; OpenAI, Anthropic, or Gemini; LangChain, LlamaIndex, or LangGraph
• Prior experience as a technical lead, staff engineer, or engineering manager for an AI/ML team

What Makes Someone a Strong Fit

We're optimizing for depth and independent judgment over years on a resume. A candidate who hasshipped a handful of production-grade AI systems and can explain the hard decisions, trade-offs, andfailures behind them is a stronger fit than one with broad but shallow exposure across many projects.

What we're looking for

• Production-level Python with deep hands-on experience in PyTorch and/or TensorFlow
• Strong experience with LLM/NLP systems: open-source and API-based models
• LLM tooling: Hugging Face Transformers; OpenAI, Anthropic, or Gemini
• MLOps platforms: MLflow, Kubeflow, Weights & Biases, or equivalent