Location: Remote, Vietnam
Employment Type: Full-time
Schedule: Flexible time
Compensation: 35,000,000 - 50,000,000 VND/month
Company: Twin Signal
About Twin Signal
Twin Signal, a ZenitechCS company, is dedicated to launching new products and expanding into new markets, building on our proven foundation of excellence. For more than 15 years, ZenitechCS has delivered reliable IT solutions to clients across California, USA. Through Twin Signal, we’re extending that legacy with focused AI-first innovation and market expansion.
Our capabilities span end-to-end technology services, including AI-enabled products, Software Development, IT Service Management, Database Administration, Information Security, Business Intelligence, Computer Networking, and Systems Administration.
Work Policy
This position is 100% remote and based in Vietnam. The Software Engineer AI Product is expected to be available and responsive during established working hours, with some flexibility based on project and operational requirements. Team members must maintain a secure working environment and prevent data exposure through cameras, bystanders, insecure networks, or other risks.
This position includes opportunities for equity compensation, allowing employees to participate in the company’s long-term growth and success.
Position Overview
This is a software engineering position. The core job is designing, building, testing, deploying, and maintaining production code - the AI/LLM components (models, RAG, agents) are one part of the stack, not a replacement for the engineering fundamentals. You're expected to write and ship real code, own services in production, and be accountable for their reliability, security, and performance, same as any backend/full-stack engineer, with the added scope of doing that around LLM-based features.
What you'll build
- User-facing and internal features powered by LLMs: chat interfaces, copilots, agentic workflows, and automation, from design through production deployment
- RAG pipelines: ingestion, chunking, embeddings, vector store indexing, retrieval, and reranking, integrated into application code
- Agent and tool-calling workflows: defining callable tools/functions, structured output contracts, and the orchestration logic (loop and graph engineering) that sequences multi-step or multi-agent tasks
- Reusable "skills": packaged prompts, tool definitions, and validation logic other features in the product can call
- Backend APIs, data pipelines, and integrations connecting your AI features to databases, vector stores, third-party APIs, and cloud infrastructure
- Secure, responsive frontend interfaces for these features, including UX for loading/confidence states and human override controls
How you'll build it responsibly
- Implement PII handling in data flows: redaction/tokenization before data hits a model or third-party API
- Implement guardrails in code: input/output validation, moderation filters, hallucination/groundedness checks, rate limits, cost caps
- Build agents with least-privilege, sandboxed tool access, and full audit logging of tool calls — the practical side of secure agent usage
- Apply standard AppSec practices to AI-specific risks: prompt injection resistance, output sanitization before rendering or downstream use, secrets scoping for tool-calling agents
- Implement human-in-the-loop (approval gates before high-risk actions) and human-on-the-loop (monitoring/alerting/kill-switch) patterns where the product calls for them, and choose the right operating mode (autonomous, supervised, copilot, batch) per feature
- Instrument features so governance and compliance teams have the lineage, logging, and audit data they need — you build the plumbing; you're not writing the policy
Full lifecycle ownership
- Prototype, build, and ship AI features end-to-end, then keep operating them: monitor drift, latency, cost, and quality in production; investigate incidents and defects; roll out model/prompt version upgrades; capture user feedback loops; deprecate features cleanly when they're retired
- Write the runbooks, eval plans, and implementation docs needed to hand off and operate what you built
Required Qualifications
- Strong professional English communication skills, both written and verbal
- 3+ years of professional software engineering experience for mid-level candidates, or 5+ years for senior-level candidates
- Proficiency in at least one general-purpose programming language such as Python, JavaScript/TypeScript, C#, Java, Go, or a comparable language
- Hands-on experience building and deploying production software products, APIs, services, or integrations
- Experience developing applications that use LLMs or other AI/ML capabilities in practical product workflows
- Strong understanding of AI application patterns such as prompt engineering, retrieval-augmented generation, embeddings, vector search, tool calling, structured outputs, and evaluation
- Experience designing and consuming REST APIs and integrating third-party systems
- Strong SQL skills and practical experience with relational databases such as PostgreSQL, SQL Server, or MySQL
- Experience building data integrations, ETL pipelines, scheduled data-processing workflows, or data-quality controls
- Experience developing web interfaces using modern frontend technologies or server-rendered web frameworks
- Familiarity with Git, code review, automated testing, CI/CD, and software release practices
- Ability to independently troubleshoot application, integration, database, data-quality, and AI-product issues
- Excellent analytical, organizational, prioritization, documentation, and collaboration skills
Preferred Qualifications
- Experience with AI frameworks and tools such as OpenAI APIs, Anthropic APIs, Azure OpenAI, AWS Bedrock, LangChain, LlamaIndex, Semantic Kernel, or equivalent
- Experience with vector databases or search platforms such as pgvector, Pinecone, Weaviate, Qdrant, Elasticsearch, or equivalent
- Experience with cloud platforms such as Azure or AWS
- Experience with data tools or frameworks such as Airflow, dbt, pandas, Spark, or equivalent
- Experience with JavaScript/TypeScript frameworks such as React, Vue, or Angular
- Familiarity with Docker, CI/CD pipelines, infrastructure-as-code, or configuration management tools
- Knowledge of secure AI development, authentication, authorization, secrets management, data privacy, and prompt-injection mitigation
- Experience with AI evaluation, observability, experiment tracking, A/B testing, or product analytics
- Experience working in an MSP, consulting, agency, SaaS, multi-client, or distributed-team environment
- Prior experience mentoring developers or leading small technical initiatives
Success Measures
- Reliable, maintainable, and timely delivery of useful AI-powered product features and services
- Accurate, secure, observable, and cost-effective AI workflows and data-processing pipelines
- Measurable improvements in AI feature quality, user outcomes, product reliability, and operational efficiency
- Reduced recurring defects, incidents, and model-quality issues through effective root-cause analysis
- Clear, complete, and current technical documentation
- Effective collaboration with clients, stakeholders, and engineering team members
- Measurable improvements in testing, deployment quality, automation, and engineering maturity
Schedule
The hours are flexible, but the candidate must be able to respond to issues during the following time: Monday to Friday: 8:00 - 17:00 (Vietnam Time).
Compensation
Software Engineer AI Product: 35,000,000 - 50,000,000 VND/month
Important Note: We will not hire a candidate who receives a “low” assessment in English proficiency.
Perks & Benefits
- End of Year Bonus: Standard bonus is valued at least one month’s salary. If the company is successful and staff have performed well, highly performing staff may receive larger bonuses. Vietnamese staff will receive bonuses before Tết.
- Health Insurance: Health insurance will be provided to all staff.
- Dental Insurance: Dental insurance will be provided to all staff.
- Vision Insurance: Vision insurance will be provided to all staff.
- Remote Work: The position is 100% remote.
- Paid Time Off (PTO): We offer a base of 15 days of paid time off in addition to 12 national holidays, for a total of 27 paid days off. Additional time off can be requested as needed.
- Mentorship: We offer on-the-job training, continuing education, code review, and technical assistance when required.