Senior Full Stack AI Engineer

Role Overview

The Senior Full Stack AI Engineer is a senior engineer who takes AI features from prototype to production end to end — owning the full stack: frontend, backend, data, deployment, and the AI layer itself including quality, evals, observability, and integration into adjacent products. This is a senior IC role with high autonomy. You will own AI initiatives, scope and de-risk new AI capabilities through prototypes, and then partner with product and other technical teams to integrate matured AI features into our products with the right safeguards. The role demands strong AI engineering fundamentals (LLMs, agents, RAG, evals, voice/multi-modal, classical ML where it fits), strong full-stack development skills, and fluency with modern AI assisted software development tools e.g. Claude Code, MCP, custom skills, agent orchestration — as the default way you work, not a side experiment.

Education

• Bachelor’s degree in Computer Science, Electrical Engineering, or a related technical discipline.
• A Master’s with an AI/ML focus is a plus.

Experience

• 5+ years of experience in AI development
• 3+ years of experience building production AI/ML systems.
• 2+ years of working as a senior contributor e.g. Senior AI Engineer, Staff AI Engineer etc.
• Track record of shipping AI features end to end — from prototype, through engineering, into production — and owning them post-launch (evals, monitoring, iteration).
• Proven full-stack delivery on at least one production product where you owned the surface area beyond just the AI/ML layer.
• Demonstrable evidence of shipping with AI tools — a GitHub repo, a working prototype, or a screen share walkthrough of something you built with Claude Code or an equivalent agentic tool.
• Experience working in a fast-paced product environment is preferred.
• Research experience at reputable venues, or patents, is a plus.
• Prior experience building products in medical billing, RCM, or adjacent U.S. healthcare verticals is a strong plus.
• Strong Python with solid software engineering fundamentals across clean code, testing, debugging, and packaging. We expect production grade engineering, not notebook only fluency.
• Deep expertise across the full LLM application lifecycle, including prompt design, RAG, fine tuning, agentic workflows, evaluation, observability, and cost and latency control.
• Hands on experience designing and shipping agentic systems using modern frameworks such as LangChain, LangGraph, CrewAI, Pydantic AI, or equivalent. You can reason clearly about when an agent is the right tool for the job and when it is overkill.
• You build evals before you ship. You know how to design eval datasets, LLM as judge pipelines, regression suites, and accuracy metrics, and you treat them as the primary mechanism for trusting AI output in production.
• Working knowledge of PyTorch, vector databases (Pinecone, pgvector, or similar), orchestration tools (Airflow, Temporal, n8n, or similar), and voice and multi modal stacks (Pipecat, Retell, ElevenLabs, Deepgram, or equivalent) where relevant. Comfortable reaching for classical ML such as rules engines, scoring models, or classifiers when those are a better fit than an LLM.
• Proven ability to take AI from prototype to production, including handling latency, cost, reliability, guardrails, fallback paths, evaluation, and monitoring.
• Strong product thinking and a clear point of view on what to build. You push back on requirements that do not improve the product, and you own outcomes rather than tasks.
• Hands-on across frontend, backend, database, APIs, and deployment. You own a feature top to bottom — UI to data model to deploy — without a hand-off.
• Working proficiency with modern frontend (React/Next.js or equivalent) and backend (FastAPI, Flask, Django, Node, or equivalent).
• Comfortable with relational and non-relational databases, caching, queues, and event-driven patterns. Can reason about data models for healthcare-grade workloads (PHI, audit trails, asymmetric read/write).
• Solid working knowledge of CI/CD (GitHub Actions or equivalent), containerization (Docker), and cloud deployment on AWS. You can stand up a production-grade pipeline without waiting for DevOps.
• Treats automated tests, logging, and monitoring as part of the definition of done — not an afterthought.

AI Fluency

• You build with AI tools, not just use them. AI is part of your daily engineering workflow — embedded in coding, code review, debugging, design, prototyping, and architectural exploration — and you expect meaningful productivity gains, not marginal ones.
• Fluent with Claude Code and equivalent agentic engineering tools (Cursor, Copilot, Windsurf, etc.). You have shipped real production code where the agent was the primary author and you were the orchestrator.
• Comfortable running multiple agents in parallel on different parts of a feature, and managing the cognitive load that comes with it. You think about software development as orchestration of a team of agents, not a typing exercise.
• Strong context engineering instincts. You know how to set up CLAUDE.md or adjacent files, custom skills, hooks, MCP servers, and project-level guidance so an agent ships code that fits the codebase from day one.
• Treats verification as the core engineering skill. You build evals, pre-commit hooks, CI gates, and review rubrics so AI output is trustworthy by default and architectural integrity is preserved as the agent does more of the typing.
• Coaches peers — inside and outside the AI Squad — on AI fluency.
• Comfortable making and defending consequential design decisions on AI features (model selection, sync vs async, batch vs streaming, model routing, caching, retrieval architecture, evaluation strategy) and writing them down.
• Designs with HIPAA, PHI handling, audit logging, encryption at rest and in transit, role-based access control, and BAA-aware data flows as defaults — not as audit checklist items.
• This is not a Tech Lead role, but you are expected to operate with very high degree of technical ownership on the AI surface area you own.

Domain Knowledge

• Familiarity with the U.S. healthcare ecosystem (medical/dental RCM, care management programs including CCM, PCM, TCM, RPM, etc.) is a strong plus.
• HIPAA-grade privacy and compliance instinct in product and AI system design.
• Awareness of EDI X12 standards (270/271, 276/277, 837P/I, 835), FHIR R4, and HL7v2 is a plus.
• Familiarity with one or more of: athenahealth, Tebra, AdvancedMD, eClinicalWorks, Epic, Practice Fusion, Waystar, Availity, Optum, Stedi is a plus. You should possess a strong ability to absorb a complex, regulated, data-rich domain quickly. You’re the kind of person who, three months in, can hold their own in a clinical workflow conversation with a billing manager.
• Excellent verbal and written English. Clear, concise communication with engineers, product, and business stakeholders.
• Documentation-first instincts. You write design docs, decision records and reviews as the primary tool for alignment, not meetings.
• Comfortable explaining complex AI trade-offs (eval results, latency budgets, model choices, failure modes) to non-technical stakeholders without losing them or oversimplifying.
• Comfortable working directly with domain experts, SMEs, and end users to gather requirements, validate ideas, and ensure alignment.
• Partners well with Tech Leads and other engineers when integrating AI features into existing products.

Cultural Fit

• Humble, collaborative, proactive, and self-driven.
• Low ego, high ownership. You raise the bar of everyone around you. You own outcomes and are comfortable jumping on calls, debugging production issues, or pairing with another engineer on a hard problem. You are not precious about where your job ends.
• Doer, not a waiter. If you see a broken flow, you find a way to fix it. You do not file a ticket and wait.
• Comfortable expanding your scope into adjacent areas — product design, prototyping — when the work calls for it
• Believes in AI-first product development and is excited to shape that culture. Thinks actively about how AI can make you and the team faster — not as a buzzword, but as something you build toward daily. You should be energized by that — not threatened — and willing to keep stretching into adjacent areas e.g. more product, more design, more strategy.
• Open to feedback and committed to continuous learning. Comfortable with ambiguity and shifting priorities.

Key Responsibilities

• Take AI features from prototype to production and own them post-launch — frontend, backend, data, deployment, AI, evals, monitoring.
• Scope, prototype, and de-risk new AI capabilities, then partner with the product squad to integrate the matured capability into the product.
• Design and ship LLM applications, agentic workflows, RAG systems, voice agents, and AI-driven automations that meet production-grade requirements (reliability, latency, cost, observability, guardrails, fallback paths).
• Build evaluation, monitoring, and guardrail infrastructure for every AI feature you ship. No AI feature reaches production without an eval suite behind it.
• Test your own work end to end. Own quality, performance, security, and reliability for the features you ship.
• Use Claude Code and equivalent agentic tools as a force multiplier across your daily workflow. Set the bar for context engineering, agent orchestration, and AI-assisted code review on the squad.
• Partner with Tech Leads and engineers in other squads to integrate AI into their products, and coach them on AI tooling.
• Work closely with product owners and stakeholders to refine requirements, validate ideas, and align technical choices with business outcomes.
• Contribute to engineering standards, code reviews, and the team’s AI engineering playbook.

Location

This is an on-site position based out of our Lahore office, with working hours from 1 PM to 10 PM PKT.