Portcast is a venture-backed, Singapore-based logistics technology startup helping freight forwarders turn data into better decisions and measurable business impact.
Portcast uses AI to surface shipment exceptions, cost risks, and the right actions, helping teams focus on the shipments that need their attention, keep freight costs under control, and make procurement decisions with confidence. Our vision is to be the AI data and intelligence layer that powers exception management and margin improvement on every shipment.
Founded in 2018 and backed by leading technology investors, we are building for an industry at a critical inflection point of digital transformation. Our team of software engineers, data scientists, and logistics experts is on a mission to build the technology that helps freight forwarders protect margin, improve operational efficiency, and differentiate their offering.
ABOUT THE ROLE:
We're looking for a Senior Machine Learning Engineer who owns solutions end to end, from framing the problem, to building and deploying the model, to running it in production. This is a hands-on, business-minded role in a lean team: you'll build ML and AI that drives real outcomes for our customers, and make pragmatic calls about cost, scale, and impact, not just model accuracy.
Our models (like predictive ETA) are one part of the picture, not the whole job. You'll own some products end to end and contribute to others, and you'll help decide what's worth building based on what moves the business. You'll work closely with our Data, Engineering, and Product teams, turning loosely defined problems, sometimes without clean baseline data, into shipped features. If you like figuring things out in ambiguity and seeing your work drive the business directly, rather than long research cycles in a big lab, you'll fit well here.
-
Own ML and AI solutions end to end: frame the problem, build the model, deploy it, and run it in production (MLOps). You own the full lifecycle, not just the modelling.
Own some products end to end and contribute to others, thinking about business outcomes, cost, and scale, and model performance.
Comfortable building from a rough outline rather than a finished spec. You'll work directly with product and customer-facing teams to turn loosely defined problems into shipped features, and re-scope quickly when priorities shift. You'll own the how, which means pushing back on a weak brief and making the call when the spec runs out. We have a strong sense of direction; the details pivot often.
Collaborate with Data, Engineering, and Product to understand the business goals behind what you are building, so what you ship actually moves the metrics that matter.
Build, test, deploy, and monitor real-time prediction models and ML algorithms that address the key business problems our product focuses on: visibility, prediction, demand forecasting, freight audit, using MLOps best practices with version control and performance tracking.
Productionise LLM-based systems where relevant, treating prompts and model behaviour as engineering artifacts.
Perform feature engineering, model tuning, and validation so models are production-ready and optimized for performance.
-
Bachelor's, Master's, or PhD in Computer Science, Engineering, or a related field.
5+ years building, deploying, and scaling machine learning models in production, with a track record of owning things end to end.
A business and software mindset over pure research. You think about business outcomes, cost, and impact, and you know when "good enough and shipped" beats a perfect model. Backgrounds that fit well include software engineers who moved into AI/ML, or ML engineers used to owning products end to end in a lean team.
Prior experience in a lean product-based startup environment: a small ML/DS team where everyone wears many hats, with no big R&D lab and no clean specs or baseline data handed to you. You create clarity and make progress without every answer upfront.
Experience in logistics tech or supply chain tech, ideally building AI-powered intelligence for freight forwarding, is a strong plus.
Experience with real-time data processing, anomaly detection, and time-series forecasting in production.
Experience with large datasets and big data technologies like Spark and Kafka.
Strong Python and SQL, with experience in cloud platforms (AWS) and containerization (Docker, Kubernetes).
Proven experience owning the full lifecycle, from R&D to production and MLOps, in fast-paced environments. You can build a model and run it in production yourself.
Hands-on experience productionising LLM-based systems. Bonus points for designing AI agents and multi-step workflows, tool/function calling, and grounding models on proprietary data through retrieval and context design, and treating prompts and model behaviour as engineering artifacts: versioning, evaluation harnesses, guardrails, and monitoring output quality, latency and cost in live systems.
First-principles thinking, a self-starter mentality, and the ability to take ownership end to end and work autonomously.
Excellent communication, with the ability to convey complex technical concepts clearly, and a strong customer-obsessed mindset. You are genuinely curious about our industry, our customers, and the logistics domain, and you let that shape what you build.
Own a high-leverage problem: your work drives real product and business outcomes, not research for its own sake.
Real ownership from day one: we're 30 people and the data science team is small, so there are no layers, nowhere to hide and everywhere to make a mark. You own ML systems end to end and grow fast.
Tech-first team: you'll work with people who care about solving hard problems with technology, and get exposure to complex product problems across software, data, ML, and logistics.
Globally distributed, remote-first flexibility: work with a fully distributed team across Asia and Europe, built on trust, accountability, and collaboration.
Curiosity: We stay close to the data and to model behaviour before we trust an output. We dig into why a model does what it does, not just whether the metric moved.
Ownership: We act like founders. We take a model from research to production and stay on it, monitoring, debugging, and improving long after it ships.
Raising the bar: We don't settle for a model that works in a notebook. We aim for systems that are reliable, scalable, and cost-aware in production.
Effective: We focus on models that create real product and customer impact, not accuracy for its own sake.
We may use artificial intelligence (AI) tools to support parts of the hiring process, such as reviewing applications, analyzing resumes, or assessing responses and identifying potential inconsistencies or verification signals in application materials based on available information. These tools assist our recruitment team but do not replace human judgment. Final hiring decisions are ultimately made by humans. If you would like more information about how your data is processed, please contact us.
