About Synthari
PRACTICAL VISION.
HONEST WORK.
We're a Singapore-based computer vision team that builds systems grounded in your data and designed for the people who will use them.
Back to HomeOur Story
HOW SYNTHARI
CAME TO BE
Synthari grew out of a frustration that felt widespread but rarely discussed: most teams exploring computer vision were being sold finished products that didn't fit their environment, or handed research-grade models with no guidance on what to do with them. Neither path led anywhere useful.
We started Synthari to fill that gap — not as a software vendor and not as an academic consultancy, but as a team that does the integration work. We train models against your actual data, define detection classes with the people who understand your product or process, and hand over systems your team can maintain and retrain as conditions evolve.
Based at Singapore Science Park, we work primarily with manufacturing, logistics, and facilities teams across the city-state. Our engagements are scoped clearly, priced transparently, and always come with documentation that doesn't require a data science background to follow.
OUR MISSION
To make computer vision genuinely accessible to operations teams — not by simplifying the technology, but by doing the hard integration work alongside you and handing over something you can actually run.
OUR APPROACH
Every engagement starts with your data, your environment, and your team's knowledge. We define scope together, build with transparency, and document every assumption — including the ones that limit performance.
OUR PROMISE
We tell you when a use case is not a good fit for computer vision. If our Scoping & Pilot engagement concludes that building a production system is not advisable, we say so — clearly, in writing.
The Team
PEOPLE BEHIND
THE WORK
A small team with complementary backgrounds in machine learning, software engineering, and industrial operations.
Neel Raghunathan
Co-Founder · Computer Vision Lead
Eight years in machine learning applied to industrial imaging. Leads model architecture, training pipelines, and calibration sessions for every client engagement.
Simin Liang
Co-Founder · Systems Integration
Background in embedded systems and factory automation. Translates model outputs into practical integration points that fit within existing operational infrastructure.
Amara Chong
Operations & Client Engagement
Manages project scoping, client communication, and delivery timelines. Former operations manager with direct experience in manufacturing quality workflows.
How We Work
QUALITY &
STANDARDS
Consistency in how we develop, document, and hand over computer vision systems.
DATA PRIVACY & HANDLING
Client image and video data is stored securely and used solely for the contracted engagement. We follow Singapore's PDPA guidelines throughout every project and delete source data upon delivery unless retention is agreed in writing.
DOCUMENTED DELIVERABLES
Every engagement produces written documentation covering model configuration, known performance boundaries, integration requirements, and retraining procedures — written for your operations team, not for other engineers.
PERFORMANCE TRANSPARENCY
We share accuracy benchmarks, false positive rates, and confidence thresholds openly during and after development. Models are delivered with honest performance notes, not optimistic summaries.
RETRAINING READINESS
Environments change. Every model we deliver comes with guidance on how to recognise performance drift and instructions for retraining with new data — so you're not dependent on us for routine maintenance.
HONEST FEASIBILITY REVIEW
Before beginning any full development project, we assess whether your data and environment genuinely support a computer vision approach. We advise against proceeding when the conditions are not right.
LOCAL ENGAGEMENT MODEL
Based at Singapore Science Park, we offer on-site visits, in-person workshops, and local project support. No offshore handoffs mid-project — the team you scope with is the team that builds.
Expertise
COMPUTER VISION
DONE WITH CARE
Object detection, visual quality inspection, and feasibility scoping are three distinct disciplines — each requiring a different combination of data preparation, model selection, and integration planning. Synthari focuses on all three because they often intersect in the same industrial environment.
Our work covers the full development arc: from initial data audit and class definition, through training, calibration, and integration, to the written handover that lets your team operate the system independently. We pay particular attention to the conditions under which models perform poorly, because that knowledge is what makes the difference between a working deployment and a shelf-bound prototype.
Singapore's manufacturing and logistics sectors present specific computer vision challenges — mixed lighting environments, variable product SKUs, multilingual signage in detection contexts, and the operational constraints of small-batch production. We've built experience in these conditions specifically because they're common to the industries we serve here.
3+
Years in Singapore
40+
Engagements Completed
100%
Deliverable Completion Rate
Work with us
READY TO EXPLORE
A COLLABORATION?
A short conversation is usually the best starting point — tell us what you're trying to see, and we'll tell you honestly what's involved.
Start a Conversation