Associate AI Engineer

Sastaticket; Karachi Division; On-site; PKR 1,674,000 - 2,567,000; Full time

Job summary

sastaticket.pk in Karachi invites applications for Associate AI Engineer to build production‑grade AI features in Python, integrate LLMs, and shape how AI is used across our platform.

You will work with the Edge Team to own components end-to-end, write clean code, review peers, and balance quality, latency, and cost in a fast‑paced travel tech environment.

This early‑career role emphasizes hands‑on ownership, collaboration across teams, and real impact on product development.

Qualifications

• Bachelor’s degree in Computer Science, Engineering, or a related field.
• 1–2 years of professional experience in software engineering or AI application development, including hands‑on experience building LLM applications.
• Strong Python fundamentals and the ability to write clean, readable, and testable code.
• Familiarity with Git, GitHub, unit testing, debugging, and code reviews.
• Understanding of REST APIs, JSON, authentication, and error handling.
• Understanding of LLM tokens, context windows, output variability, and quality, latency, and cost trade‑offs.
• Experience writing and refining prompts using clear instructions, examples, and structured outputs.
• Familiarity with tool calling and agent workflows, including control flow, state management, and error handling.
• Exposure to an AI framework such as LangChain, LangGraph, or CrewAI, or experience building directly with model APIs.
• Understanding of LLM evaluation, including deterministic checks, LLM-as‑a‑judge approaches, and regression evaluation.
• Understanding of how logs and traces help investigate errors, unexpected outputs, latency, token usage, and cost.
• Ability to prepare evaluation datasets covering typical scenarios, edge cases, and known failures.
• Willingness to ask questions, communicate clearly, receive feedback, and learn.

Responsibilities

• Build and maintain AI-powered product features, primarily in Python, with guidance from experienced engineers.
• Write clean, readable, testable code and participate in code reviews, debugging, and engineering discussions.
• Integrate LLMs and AI tools into applications using model APIs and frameworks like LangChain, LangGraph, or CrewAI.
• Design and refine prompts using clear instructions, examples, and structured outputs.
• Contribute to tool-calling and agent workflows, including control flow, state management, and error handling.
• Evaluate LLM behaviour using deterministic checks, LLM-as-a-judge approaches, regression evaluation, and practical evaluation datasets.
• Prepare evaluation datasets covering typical scenarios, edge cases, and known failures.
• Use logs and traces to investigate errors, unexpected outputs, latency, token usage, and cost.
• Work with REST APIs, JSON, authentication, and error handling when integrating services and AI capabilities.
• Collaborate with engineers and cross-functional colleagues to understand problems, develop features, and improve product quality.
• Use AI coding assistants thoughtfully, reviewing and validating generated code rather than treating generated output as automatically correct.
• Take greater ownership and work more independently as your capability and judgment grow.

Skills

Python programming

LLM concepts

Prompt writing

REST APIs

Software engineering

Communication

Education

Bachelor’s degree in Computer Science, Engineering, or a related field

Tools

Git

GitHub

Job description

Welcome, You are about to start your application process for

Hiring for Associate AI Engineer at sastaticket.pk

Build AI products that turn intelligence into real business impact
help shape how AI is used across sastaticket.pk by building reliable, production-ready applications that solve real business problems — combining Python, Large Language Models, and modern AI tooling to automate intelligently, improve customer and team experiences, and turn emerging technology into measurable value.

About sastaticket.pk

sastaticket.pk is Pakistan’s leading technology-driven travel platforms, helping millions of travellers plan and book journeys with greater ease, confidence, and choice. We are building a profitable, ambitious, engineering-led company that solves meaningful problems across travel, payments, pricing, customer experience, automation, and marketplace reliability.
We believe great businesses are built by great people. We give talented individuals meaningful ownership, exposure to important decisions, and the opportunity to grow through real responsibility and measurable impact.

Why This Role Exists

AI can create meaningful business value only when it is translated into reliable products, evaluated carefully, and improved based on real-world performance. This role exists to help the Edge Team build AI-powered applications that solve practical business problems while strengthening the engineering quality, evaluation discipline, and observability needed to use LLMs responsibly in production.

About the Opportunity

As an Associate AI Engineer in the Edge Team, you will work alongside experienced engineers and contribute to real AI-powered product features. You will work primarily in Python, integrate LLMs and AI tools, evaluate model behaviour, debug issues, and collaborate across functions. This is an early-career role designed for someone who already has hands‑on exposure to software engineering or AI application development and wants to grow through real product ownership, feedback, and increasingly independent execution.

What You’ll Do

• Build and maintain AI-powered product features, working primarily in Python with guidance from experienced engineers.
• Write clean, readable, testable code and participate actively in code reviews, debugging, and engineering discussions.
• Integrate Large Language Models and other AI tools into applications using model APIs and/or frameworks such as LangChain, LangGraph, or CrewAI.
• Design and refine prompts using clear instructions, examples, and structured outputs.
• Contribute to tool-calling and agent workflows, including control flow, state management, and error handling.
• Evaluate LLM behaviour using deterministic checks, LLM-as-a-judge approaches, regression evaluation, and practical evaluation datasets.
• Prepare evaluation datasets covering typical scenarios, edge cases, and known failures.
• Use logs and traces to investigate errors, unexpected outputs, latency, token usage, and cost.
• Work with REST APIs, JSON, authentication, and error handling when integrating services and AI capabilities.
• Collaborate with engineers and cross-functional colleagues to understand problems, develop features, and improve product quality.
• Use AI coding assistants thoughtfully, reviewing and validating generated code rather than treating generated output as automatically correct.
• Take greater ownership and work more independently as your capability and judgment grow.

What Success Looks Like

• You contribute production-quality AI features that are understandable, testable, and maintainable by the team.
• You can explain how an LLM-powered feature works, where it may fail, and how its quality is being evaluated.
• Prompts, integrations, and agent workflows become more reliable through testing, evaluation, and iteration.
• You use logs, traces, and evaluation results to investigate problems rather than relying only on manual observation.
• You understand and consider quality, latency, token usage, and cost trade‑offs when making implementation decisions.
• You ask useful questions, act on feedback, communicate clearly, and require less supervision as you grow.
• You use AI tools to increase leverage while maintaining engineering judgment, validation, and accountability.

What You’ll Learn

• How production LLM applications are designed, integrated, evaluated, monitored, and improved.
• Practical prompt engineering, tool calling, agent workflows, and model API integration.
• Evaluation design, regression testing, observability, and failure analysis for non‑deterministic systems.
• Software‑engineering practices for reliable AI applications, including testing, debugging, code review, and API design.
• How AI products are applied to real business and customer problems inside a technology‑enabled travel company.

At sastaticket.pk , growth isn’t determined by tenure — it’s earned through ownership, capability, learning speed, and measurable impact. As you grow, you’ll take greater ownership of AI features and technical decisions, work more independently, deepen your expertise in production AI systems, and contribute to increasingly complex products and workflows.

What We’re Looking For

• Bachelor’s degree in Computer Science, Engineering, or a related field.
• 1–2 years of professional experience in software engineering or AI application development, including hands‑on experience building LLM applications.
• Strong Python fundamentals and the ability to write clean, readable, and testable code.
• Familiarity with Git, GitHub, unit testing, debugging, and code reviews.
• Understanding of REST APIs, JSON, authentication, and error handling.
• Understanding of LLM tokens, context windows, output variability, and quality, latency, and cost trade‑offs.
• Experience writing and refining prompts using clear instructions, examples, and structured outputs.
• Familiarity with tool calling and agent workflows, including control flow, state management, and error handling.
• Exposure to an AI framework such as LangChain, LangGraph, or CrewAI, or experience building directly with model APIs.
• Understanding of LLM evaluation, including deterministic checks, LLM-as‑a‑judge approaches, and regression evaluation.
• Understanding of how logs and traces help investigate errors, unexpected outputs, latency, token usage, and cost.
• Ability to prepare evaluation datasets covering typical scenarios, edge cases, and known failures.
• Willingness to ask questions, communicate clearly, receive feedback, and learn.

Nice to Have

• Personal projects or open-source contributions involving LLM applications.
• Familiarity with RAG, embeddings, vector search, BM25, or hybrid retrieval.
• Experience with open‑weight or locally hosted LLMs using tools such as Ollama, or model integrations through OpenRouter or LiteLLM.
• Exposure to monitoring tools such as LangSmith, Langfuse, MLflow, or OpenTelemetry.
• Experience integrating automated evaluations into CI workflows.
• Familiarity with cloud platforms such as AWS or GCP.
• Experience using AI coding assistants and reviewing the code they generate.
• Exposure to transcription, voice agents, or real‑time conversational workflows.
• Experience working with customer support teams or evaluating service quality.
• Interest in experimenting with AI and applying lessons from documentation, technical articles, or research.
• Build trust by being transparent about what they know, what they do not know, model limitations, risks, and mistakes.
• Take extreme ownership by following work through from implementation to testing, evaluation, debugging, and improvement.
• Stay customer-focused by connecting AI work to useful, reliable experiences and real business problems.
• Learn continuously by experimenting, reading, asking questions, acting on feedback, and applying new knowledge quickly.
• Use judgment under ambiguity, especially when model behaviour is variable or requirements are incomplete.
• Deliver with speed and quality by moving quickly without treating generated code or model output as automatically correct.
• Use AI and modern tools to increase leverage while maintaining validation, security, and engineering judgment.
• Collaborate with low ego, communicate clearly, and make the people and systems around them stronger.

What We Offer

• Medical coverage for self and dependents.
• Gratuity.
• Annual travel allowance.
• Continuous learning and professional development opportunities.
• Exposure to senior leadership and important business decisions.
• Meaningful ownership and career growth based on performance and impact.
• A collaborative, ambitious, and high‑performance work environment.