Software Engineer, Algorithms Team

The 4M story is likely one you haven't heard before: We are on a mission to reveal the world below us — to do for the world below ground what Google Maps did for the world above. By leveraging cutting-edge technology, we are mapping the subsurface infrastructure to make reliable, real-time utility data accessible to the construction industry — completely transforming a traditional industry. We're a growing startup with 100 employees currently based in Tel Aviv, Israel, and Austin, Texas.

At the core of this mission is a computer-vision engine that detects and localizes utility infrastructure from millions of diverse types of images. That engine only creates value if it runs reliably at scale — and that is where this role lives. We are looking for a builder: an engineer who is genuinely energized by the challenge of making complex systems work, who plans like an architect, debugs relentlessly, and learns whatever the problem demands.

The Opportunity

We are looking for a Software Engineer to join our core Algorithm team as an dedicated data-core / platform engineer. You will support the algorithm engineers developing detection algorithms to design and build the infrastructure they run on: the object-detection (OD) pipeline, its orchestration and deployment, its data and model catalog, and the benchmark and labeling systems that drive model improvement. This is a high-autonomy role with real, end-to-end ownership of production systems from day one.

The technologies listed in this description are examples of our day-to-day — not a rigid checklist. We care far more about how you think, how you plan, how you debug, and how fast you learn than about which specific tools you have already used. If you love the craft of making things work and want to go deep, you can learn the rest here.

In the AI era, being a strong engineer means more than writing great code. It means operating as an architect who directs AI agents — designing the solution with clarity, then guiding them to execute it at a level and speed that wasn't possible before. We are building a culture where this is the norm, and we're looking for someone who is excited to work and grow in that direction.

What You'll Do

• Take on hard, open-ended infrastructure challenges and make them work — designing, building, decoupling, and hardening the systems behind our object-detection pipeline, from data and model management to benchmarking, so everything runs reliably at scale.
• Architect and build the backbone of the OD pipeline — orchestration (Airflow on Kubernetes), data plumbing (S3 / PostGIS / SQS), CI/CD, and deployment across multiple environments — designing clean interfaces and data contracts the algorithm team can build on with confidence.
• Debug across the whole stack, wherever the problem leads — a stuck DAG, a flaky pipeline stage, a slow query, a GPU/driver mismatch — and turn one-off firefights into lasting fixes and better observability.
• Own the data and model lifecycle: versioned datasets and model weights with clear provenance, and the labeling → export → retraining loop that keeps the models improving.
• Learn fast and go deep. Pick up new tools and new layers of the stack as the work requires, and raise the team's engineering and operational standards as you go.
• Partner closely with algorithm engineers and the data-collection / labeling operations team to turn research prototypes into robust, scalable production systems.
• Integrate AI tools into your workflow and grow into operating as an architect who directs AI agents — designing the solution, then guiding them to build it.
  
  

Who You Are

This is what matters most to us:

• You love the challenge of building. You're energized by taking something messy, broken, or undefined and turning it into a system that runs reliably. You get real satisfaction from making things work.
• You're an architect by nature. You think in systems, plan before you build, and design clean boundaries and interfaces rather than bolting things on.
• You're a strong debugger. You're methodical and curious, and you enjoy chasing a problem through every layer — code, data, infrastructure, hardware — until you find the root cause.
• You're eager to learn. You don't need to already know our whole stack. You need to be the kind of person who masters new things quickly and genuinely enjoys doing so.
  
  

Requirements:

Required Qualifications

• B.Sc. in CS, EE, or a related field, with 4+ years of professional software engineering experience.
• Strong Python and software-engineering fundamentals, with a high bar for clean, production-grade, well-tested code — whether you write it by hand or direct AI agents to produce it (our stack is Python 3.13).
• Real experience building and operating production systems end-to-end (backend, data, platform, or infrastructure) — not just shipping features on top of someone else's system.
• Comfort with cloud infrastructure and relational databases (we use AWS and PostgreSQL/PostGIS).
• Demonstrated ability to design systems and to debug hard problems — the two aptitudes at the heart of this role.
• Good communication — works well across disciplines with algorithm and operations teams.
  
  

Our Stack — What You'll Work In and Grow Into

Deep prior experience with all of these is NOT required. We expect a strong engineer to have worked with some of them and to be excited to master the rest here:

• Workflow orchestration — Airflow (or an equivalent such as Prefect, Dagster, or Argo).
• AWS — S3, SQS, ECR, EC2/EKS, EFS.
• Containers and deployment — Docker, Kubernetes, KEDA/Karpenter auto-scaling.
• CI/CD — GitHub Actions or equivalent.
• Data layer — PostgreSQL/PostGIS, schema design and migrations, S3/EFS storage.
• Secrets, monitoring, and reliability — Vault, Datadog, Slack alerting.
  
  

AI-First Mindset

We are building an AI-first engineering culture, and we're looking for engineers who are genuinely excited about working this way. You don't need to arrive with a fully formed AI workflow — but you should have the curiosity and the drive to develop one.

In practice, this means:

• You're already experimenting with AI tools in your work — code assistants, LLMs, or AI-augmented workflows — and you want to go further (e.g. using agents to accelerate refactors, infrastructure-as-code, debugging, and test-writing).
• You're open to learning how to operate as an architect directing AI agents: holding a clear picture of the system, then guiding them to get there.
• You see AI not as a threat to engineering craft, but as a multiplier of it.
  
  

Specific tools and methods can be learned together. What matters is the mindset and the motivation to grow into this way of working.

Preferred Qualifications (not required!)

• GPU / CUDA and model acceleration (e.g. TensorRT, quantization) for faster inference.
• MLOps or ML-platform experience — model registries, benchmark/eval pipelines, inference infrastructure.
• Kubernetes in production; event-driven auto-scaling (KEDA) / node provisioning (Karpenter).
• Track record of decoupling a monolith into services and defining clean data contracts.
• Familiarity with annotation tooling (CVAT) or human-in-the-loop data pipelines.
• PostGIS / geospatial data; infrastructure-as-code (Terraform).
  
  

Why Join 4M

If you love the craft of making complex systems work — and want your work to directly enable a product that maps the world below ground — this is that role.

• Real ownership from day one — a genuine, high-impact first mission, not a ramp-up exercise.
• High-leverage work: the platform you build is what the entire algorithm team ships on.
• Endless hard, satisfying problems to make work — architecture, debugging, scaling, and reliability.
• You'll grow into operating as an architect who directs AI agents — a scope of influence that compounds over time.
• High-autonomy culture that trusts engineers to lead — real responsibility, no micromanagement.
• Pre-scale startup (~100 people): your influence on the technical culture, the stack, and the direction is real and lasting.