AI & Agent Development Engineer (Remote)

AI summary

Tongston is hiring an AI & Agent Development Engineer for a remote contract role. The role involves building AI workflows, integrating LLMs, and working with AWS and various technologies. Ideal for early-career AI engineers.

• Remote contract position
• Requires experience with LLM orchestration, AWS, and full-stack development
• Work with open-source and closed-model APIs

Original source description

AI & Agent Development Engineer (Remote) at Tongston ⏲ Jul 28, 2026, 9:05 AM ⋕ View all Data Science & Analytics jobs Tongston is a multi-award-winning brand providing entrepreneurial education, media, enterprise & finance services for sustainable socio-economic development. T-World is Tongston's AI powered digital entrepreneurial thinking ecosystem providing integrated media, enterprise, finance and entrepreneurial education: services to K-12 Students, HE Students, Entrepreneurs, Intrapreneurs and their institutions globally to become Valuable, Influential & Profitable. We are recruiting to fill the position below: Job Title: AI & Agent Development Engineer Location: Remote Employment Type: Contract Overview Tongston is seeking an AI & Agent Development Engineer to translate Tongston’s AI strategy, governance requirements, product specifications, KB logic, user-context logic, and model-routing decisions into production-ready AI workflows within a platform. This role is ideal for an early-career AI engineering professional who can help build, integrate, test, and improve AI-powered product features across the stack, including agent workflows, LLM-enabled experiences, retrieval logic, APIs, and supporting backend systems. This role is not limited to embedding third-party AI APIs. It requires practical understanding of LLM orchestration, open-source model selection, multi-model workflow design, retrieval systems, compute infrastructure, AWS-based deployment considerations, cost optimisation, and secure integration into T-World’s existing GitHub codebase and platform architecture. It will operationalize the AI layer by combining approved open-source LLMs, selected closed-model APIs where justified, internal knowledge-base logic, user data, platform permissions, and governed workflow orchestration into reliable AI-enabled product features. You will work closely with education, AI, data/research/economics team members; and a cross functional IT team of UX/UI designers, backend & front-end engineers, and other key team members/stakeholders. You will participate in Agile ceremonies including planning, development, testing, and iteration. Core Stack: Engineering & LLM Orchestration: Multi-model / multi-layer LLM orchestration Open-source LLM integration and evaluation Dynamic function invocation using OpenAI / GPT-wrapper or equivalent where approved Flowise, Dify, LangChain, LangGraph, LlamaIndex, or similar orchestration frameworks API-based LLM workflows Back-End: API services and serverless routes where applicable Storage: AWS S3 including S3, Lambda, Step Functions, State Machines, Bedrock where applicable, and related cloud services Authentication: Firebase (synced with MongoDB) MongoDB (Atlas & Compass) CI/CD: GitHub Actions Node.js with Express (API services / serverless routes where applicable) Front-End: Reactjs, Typescript, Tanstack-Router/Querry, Zustand Roles and Responsibilities: Implement retrieval and knowledge-base integration patterns, including storage and access logic for structured and semi-structured content, references, metadata, tagging logic, permissions, and linked evidence. Build, test, and optimize agent workflows using platforms such as Flowise, Dify, LangChain, LangGraph, LlamaIndex, or other approved orchestration tools, while supporting movement beyond no-code or low-code setups where greater control, customization, maintainability, or scale is required. Ensure all AI-enabled features are production-ready, reliable, cost-conscious, secure, maintainable, and aligned with Tongston’s long-term goal of building T-World’s AI layer. Build, implement, improve, and operationalize AI agents, assistant workflows, retrieval flows, and LLM-powered product features for T-World using approved tools, frameworks, orchestration platforms, APIs, open-source models, and code-based approaches as appropriate. Evaluate and recommend suitable LLMs or AI components for specific tasks, considering model capability, cost, latency, licensing, privacy, reliability, deployment feasibility, and whether the model should be used for free/basic functionality or premium/pay-per-use functionality. Log AI requests, model calls, retrieval events, outputs, failures, fallback events, premium-model usage, and relevant debugging information for review, monitoring, cost control, and future improvement. Escalate major technical, product, model, integration, compute, reliability, cost, privacy, or architectural risks to management in a timely manner. Work with substantive reviewers, QA colleagues, AI Governance, Education, Data, Back End, Front End, and UI/UX teams to ensure that outputs are technically functional, contextually appropriate, governed, usable, operationally fit for purpose. Support latency management, loading states, fallback messages, retries, error handling, degraded-mode behaviour, and resilience patterns for AI-enabled features. Work closely with backend, frontend, and full-stack engineers to ensure clean AI integration across UI → API → AI orchestration → retrieval → database → output. Implement technical logic that allows T-World AI to suggest, prefill, personalize, validate, or generate outputs across the platform, ensuring that outputs respect backend permissions, visibility rules, approved data-source rules, moderation states, and restricted-content controls. Implement workflows that reduce unnecessary dependency on major closed-model providers, while still allowing approved premium model usage where advanced reasoning, accuracy, or performance justifies the cost. Support the development of T-World’s proprietary multi-layer AI model by combining approved open-source LLMs, embeddings, retrieval systems, internal KB structures, user-context logic, and task-specific model routing into governed and reusable AI workflows. Maintain technical documentation covering model choices, orchestration workflows, API dependencies, AWS/compute assumptions, retrieval logic, testing outcomes, implementation decisions, risks, and handover notes. Carry out additional tasks as required by management in support of product goals & req’s. Implement application logic and structured data flows in MongoDB and related systems to support AI-enabled use cases, including retrieval, personalization, content mapping, user-specific recommendations, traceable workflow behaviour, and analytics-linked AI outputs. Enforce backend permissions, visibility rules, moderation states, and KB approval status across all AI outputs. Translate product, user, AI governance, KB, and platform requirements: into working technical flows, including user input handling, prompt orchestration, model routing, retrieval logic, guardrails, state handling, tool/function calling, fallback behaviour, and output delivery. Write clean, maintainable, secure, and well-documented code, ensuring clear separation across API, AI orchestration, model routing, retrieval, data, logging, and platform-integration layers. Handle weak input, no results, AI failure, timeouts, partial retrieval, partial model failure, and insufficient-context scenarios with clear fallback behaviour. Implement backend services, APIs, retrieval layers, storage logic, and integration logic required to power AI features across web and mobile product experiences, while aligning closely with the wider engineering architecture. Incorporate AI-related code into the existing T-World GitHub codebase in a clean, reviewable, maintainable, and modular manner, using appropriate branches, pull requests, documentation, and handover notes. Work with compute and cloud infrastructure considerations, particularly AWS, to ensure AI workflows can run, scale, store data, retrieve data, log activity, and handle latency or failure in a production-oriented environment. Support AI-linked platform behaviour such as premium AI workflows, free-tier AI workflows, model fallback, usage limits, cost tracking, and escalation to higher-cost models only where justified. Collaborate with other engineers on AI-enabled experiences in the app, including AI-powered assistants, search, recommendations, autofill, content generation, workflow suggestions, and widget-to-widget AI journeys. Design and execute testing for AI systems, including weak or empty input, no-result scenarios, AI failure or timeout, model fallback, cost-trigger thresholds, consistency, permission enforcement, and workflow reliability. Support deployment, release readiness, monitoring, and iterative improvement of AI-enabled features in live or staging environments, including coordination on versioning, configuration, testing, observability, and stable release cycles. Production-Readiness Expectations All AI-enabled features built under this role must be suitable for real product environments, not just prototypes. This includes: Ensuring: AI does not bypass backend validation or permissions Rejected or restricted content never appears in outputs Only approved and permitted data is used in outputs Clear handling of: AI failure or timeout No-result retrieval scenarios Partial system failure (e.g., retrieval works but AI fails) Weak or insufficient input Designing for: Real-world latency and user experience: constraints Maintainability and handover Traceability and debugging Implementing: Retry or degradation strategies where applicable Dtructured logging for AI requests, failures, and outputs Fallback behaviour where AI confidence is low In addition, the role must ensure that AI-enabled features are designed for: The role is expected to think beyond “the AI works” and ensure that the AI system is reliable, safe, scalable, cost-conscious, maintainable, and strategically owned by Tongston as much as possible. Model-routing control open-source and closed-model separation cost optimisation and premium-use logic compute and cloud infrastructure constraints AWS deployment readiness where applicable secure retrieval and storage GitHub-based maintainability traceability of model calls and outputs clear fallback when cheaper/free models are insufficient future extensibility toward a more proprietary T-World AI layer Knowledge, Skills & Attitude Requirements: Experience: Level: This role is most ideal for early-career engineers who has hands-on experience: building backend systems and has also begun working practically with LLMs, AI agents, retrieval workflows, or AI-enabled product features. Relevant experience: may include personal projects, internships, volunteering, open-source work, hackathons, freelance engagements, research projects, or professional roles involving AI workflows, backend engineering, agent integration, API integration, cloud infrastructure, or product implementation. The ideal candidate does not need to have trained a foundation model from scratch, but must understand how modern AI systems are practically assembled using APIs, open-source models, orchestration frameworks, retrieval systems, user-context logic, cloud infrastructure, and product code. Technical Knowledge & Skills Ability to contribute AI-related code into the existing T-World GitHub codebase using branches, pull requests, review cycles, documentation, issue tracking, and release workflows, while avoiding disconnected prototypes. Familiarity with authentication, Firebase Authentication, secure access patterns, backend visibility rules, moderation states, approved data-source rules, and how to ensure AI outputs do not bypass platform permissions or backend validation. Working knowledge of Node.js, Express, REST APIs, third-party API integration, internal service integration, and server-side logic required to connect user actions, platform data, AI workflows, storage, authentication, and output delivery. Working knowledge of MongoDB, MongoDB Atlas, and MongoDB Compass, including schemas, collections, query logic, data validation, debugging, and structuring/retrieving data for AI-enabled use cases. Experience: or strong practical exposure to LLM APIs, agent workflows, prompt orchestration, retrieval logic, tool/function calling, model routing, fallback logic, and structured AI integrations for product features such as suggestions, autofill, personalisation, recommendations, search, and content generation. Degree, HND/ND, current study, or equivalent practical experience: in Computer Science, Software Engineering, Data/Analytics, Information Systems, Mathematics, Statistics, Engineering, AI/ML, or related fields. Working awareness of open-source/lightweight LLMs, small language models, domain-specific models, embedding models, rerankers, and model repositories such as Hugging Face; with ability to assess models based on task fit, accuracy, latency, cost, licensing, privacy, deployment feasibility, and integration constraints. Ability to test and troubleshoot AI workflows across UI → API → AI orchestration → retrieval → database → output flows, including weak input, no-result retrieval, timeout, failure, fallback behaviour, logging, monitoring, traceability, maintainable code, and handover documentation Ability to support Tongston’s goal of building a proprietary T-World AI layer by combining approved models, retrieval systems, internal KB logic, user-context logic, permissions, and governed orchestration flows, rather than relying only on major closed LLM providers. Familiarity with AWS and AI-related infrastructure, including S3, Lambda, Step Functions, State Machines, Bedrock where applicable, storage, logging, scaling, latency, cost, and the compute infrastructure needed to run, call, host, or orchestrate AI/LLM workflows. Soft Skills & Attitude: Curious, proactive, and committed to continuous learning Strong alignment with Tongston’s mission, values, and culture. Learn more about Tongston’s culture: https://tongston.com/about Ability to prioritize and execute tasks independently in a remote environment to schedule Strong analytical and problem-solving skills Clear written and verbal communication High attention to detail and ownership mindset Nice-to-Have Knowledge & Skills Experience: with model evaluation, prompt testing, AI workflow testing, red-teaming, fallback testing, or quality assurance for AI outputs. Experience: with vector databases, embeddings, semantic search workflows, reranking, chunking, metadata filtering, and retrieval evaluation. Experience: building internal tools, admin panels, or workflow automation that support KB maintenance, QA, review, analytics, reporting, or operational oversight. Ability to collaborate with front-end engineers on AI-linked UI behaviour, including prompts, suggestion panels, autofill options, loading states, fallback messages, error states, review/edit flows, and generated-output displays. Experience: with LangChain, LangGraph, LlamaIndex, Flowise, Dify, or similar orchestration frameworks. Basic Web3 exposure where relevant to wallets, signing flows, token interactions/conversion, and on-chain vs off-chain boundaries. Exposure to CI/CD, GitHub Actions, deployment pipelines, staging environments, and production support practices. Experience: with AI-powered search, assistants, recommendations, content generation, autofill, summarisation, classification, or related AI-enabled product features. Salary and Benefits: Eligibility for discretionary performance-based bonus/allocation Up to 10 working days of approved paid time away from active delivery annually. Exposure to user research, product strategy, user journeys, wireframing, prototyping, usability testing, analytics-informed design improvement, and cross-functional product delivery across web, mobile and other digital platforms. A monthly Remote Productivity Support Package. Complimentary premium access to T-World, Tongston’s integrated digital platform. Team-based execution and support, including an overall project management structure, reporting to a line manager, and collaboration within a sprint-based team workflow with defined weekly deliverables. Participation in Tongston Academy sessions (scheduled periodically): Sales, communication & networking Personal finance & investing Business development & strategy Project management, Cybersecurity awareness & Digital automation tools. Personal branding (includes CV/LinkedIn review) Monthly professional fee of NGN175,000. A structured learning and growth pathway supported through twice-yearly performance review, development and growth conversations (January & July). If you join after a review cycle, your first review will take place in the next review window. The review helps you document and communicate how you’ve grown, developed and demonstrated being Valuable, Influential & Profitable. Professional Visibility: listed on Tongston’s website. A company-provided laptop and access to Tongston-funded platforms, software, licensed tools and other approved resources required to perform your role effectively. Eligible for the Tongston Annual Personnel High Performance Award which comes with cash prizes and certificates you can include in your profile for future jobs. Pathway to full-time employment opportunity to secure additional benefits, including pension, HMO, statutory employee leave categories, promotion-based entitlements, and staff welfare benefits. See our Careers page for more info. Learning & Development budget: NGN200,000 to support continuous upskilling, role-specific learning and broader professional development. Hpw to Apply Interested and qualified candidates should: Click here to apply online Important Information and Notice Applications will be reviewed on a rolling basis, so we strongly encourage early submission. Where you can provide information about yourself, upload your CV & portfolio where applicable, and answer some basic pre-screening questions that demonstrate how you meet the knowledge, skills and attitude requirements: for the role, in line with the JD. If you have any questions or clarifications, please write to us at: [email protected] with [email protected] in copy. Please use the subject “IC Application for - [name of the role]” in directing any correspondences to us. If shortlisted, you will progress through the rest of the recruitment process - technical assessment, personality & multiple intelligence check, conversation with the line manager and one or more team members, reference checks, and, if successful, an offer & if accepted, onboarding. Although this is a remote opportunity, candidates based in Gombe, Bauchi, Ilorin, Lagos or Abuja will be viewed favourably. For candidates who have not yet completed their NYSC, we may also be able to consider an NYSC engagement, particularly for those based in Gombe, Lagos or Abuja. Application Deadline: 31st August, 2026. View all Jobs in Nigeria Remote