The posting
About Sigmawave AI
Sigmawave AI builds visual data and computer vision systems that work in the real world. We create synthetic and real-world visual datasets to train computer vision models, and we deliver computer vision projects end to end: data capture and generation, model development, and deployment on edge and server hardware.
Our systems run in live operational environments where cameras, lighting, hardware limits and client acceptance criteria all matter. We are a SUSE Linux Ready partner, and we work with multiple government agencies, government-linked corporations (GLCs) and major technology companies such as HPE and Dell.
The role
You will own the technical direction and delivery of Sigmawave AI's engineering work across its three branches: Computer Vision, Synthetic Data Generation, and Software Architecture and Planning.
This is a hands-on leadership role. You set the technical approach, make the architecture calls, plan the work, lead the engineers who do it, and stay close enough to the code to unblock them. On client projects you are the technical authority: the person the business relies on to turn requirements into working systems that ship on time.
What you will own: Computer Vision
You will lead the technical direction of our Video Management System (VMS): a centralised platform for live video monitoring, recording playback and video event management, with built-in Vision AI operations to deploy, monitor and validate computer vision models across large, multi-site environments.
The VMS lets operators manage the full model lifecycle, from validation and deployment to anomaly detection and incident resolution, in one platform and without relying on vendors or data scientists. It cuts the need for lean operator teams to watch live feeds by hand.
- Lead model development across the problems we solve: object detection, segmentation, tracking, classification, re-identification, pose estimation, video analytics and multi-camera systems.
- Set the standard for how we train and evaluate: reproducible experiments, benchmark datasets, metrics tied to client acceptance criteria, and honest reporting of what a model can and cannot do.
- Own deployment and optimisation: ONNX and TensorRT export, quantisation, inference on edge devices and GPU servers, and the throughput and latency budgets that go with them.
- Put MLOps in place: training pipelines, model registry, data and model versioning, production monitoring, and a retraining loop fed by field data.
- Review the team's model work, pair with engineers on the hard problems, and bring in new techniques when they earn their place.
What you will own: Synthetic Data Generation
- Own the synthetic data pipeline end to end: scene and asset creation, rendering and simulation, procedural variation and domain randomisation, automatic ground-truth labelling, dataset QA and versioning.
- Decide when synthetic data, real-world capture, or a blend is the right answer for a given model, and measure the sim-to-real gap so the data demonstrably improves model metrics.
- Run real-world capture and annotation programmes with the same rigour: capture plans, annotation guidelines, quality gates, data governance and client confidentiality.
- Build the tooling that makes generation repeatable: configuration-driven pipelines, automated validation, dataset statistics and coverage reports.
- Keep the data stack current across rendering engines, generative image and video models, and labelling tools; make the build-versus-buy calls.
What you will own: Software Architecture and Planning
- Design the architecture for our products and client systems: video ingestion, inference services, APIs, storage, dashboards, integration with client systems, and deployment on-premise and in the cloud (Linux, containers, Kubernetes, GPU scheduling).
- Turn client requirements and proposals into technical plans: scope, work breakdown, estimates, milestones, dependencies and risks. Keep the plan honest as the project runs.
- Set engineering standards: code review, branching and release process, testing with the QA team, documentation, security and data handling.
- Own the technical roadmap: what we build once and reuse across projects, what we retire, and where we invest in platform versus project work.
- Make the tooling and build-versus-buy decisions, and keep development and deployment environments reliable.
Leading the team and delivering for clients
- Lead the engineers across all three branches, in Singapore and remote, including contractors and interns: set direction, review work, unblock, mentor, and run performance and growth conversations.
- Hire: define roles, interview, and build a team that balances research depth with delivery discipline.
- Be the technical face to clients and partners: requirements workshops, design reviews, demos, the technical sections of proposals and tenders, and acceptance.
- Prioritise across concurrent projects and protect delivery dates; raise it early when scope, timeline and capacity do not fit.
- Work closely with QA, project management and the business on what ships and when.
What we are looking for
- Strong in machine learning and software engineering, with 8+ years of hands-on experience, including 3+ years leading engineers or a technical team.
- Computer vision systems shipped to production, not only prototypes, including deployment on edge or on-premise hardware.
- Strong Python and C++ (or an equivalent systems language); PyTorch or TensorFlow; OpenCV; ONNX, TensorRT or a comparable deployment toolchain.
- Hands-on experience with synthetic data or simulation for machine learning (3D rendering, procedural generation or generative models), and with real-world dataset creation and annotation.
- Strong in system architecture design: distributed systems, APIs, containers and Kubernetes, Linux, CI/CD and GPU infrastructure.
- A track record of planning and delivering multi-month projects on time, with clear estimates and risk management.
- Clear communication with clients and non-technical stakeholders; comfortable owning technical decisions and explaining them.
- Singapore local (Citizen or Permanent Resident), based in Singapore or willing to relocate.
Good to have
- Participated in and delivered enterprise-grade or public sector projects.
- Multi-camera video analytics at scale, including tracking and re-identification across cameras.
- NVIDIA Jetson, DeepStream or Triton; model compression and quantisation.
- Omniverse or Isaac Sim; neural rendering (NeRF, Gaussian splatting) or diffusion models for data synthesis.
- Enterprise Linux (SUSE, RHEL) and on-premise GPU clusters.
- Writing the technical sections of tenders and proposals.
- Building and mentoring teams across countries.
What we offer
- Real ownership across data, models and systems, not one slice of the stack.
- A small, senior team where your technical decisions ship, and clients who measure results in the field.
- Broad exposure to the corporate and software world, working on multi-agency projects with government agencies, GLCs and major technology companies such as HPE and Dell.
- Direct influence on the company's technical roadmap and on how the engineering team grows.
- Salary commensurate with experience.
How to apply
Send your CV and a short note on one computer vision system you took to production, what it did, and what you would do differently, to [email protected]
