Senior AI Developer (C++ / Gameplay AI)

Job Overview

We’re building a next-generation AI stack for large-scale, session-based vehicular combat. You’ll own end-to-end AI features—from behavior technical design to integration with our server simulation, sensory systems, telemetry, and ML inference services. You’ll help us move from “default bots” to adaptive, performant, debuggable opponents and squadmates that feel human without breaking fairness or the server budget.

What will you do?

• Design & Implement Behaviors: Ship production-grade AI using GOAP, Utility AI, and/or BTs, choosing the right tool per feature (target selection, positioning, driving, firing, evasion, cooperation).
• Author Sensors & World State: Build reusable sensors, perception pipelines, blackboards/knowledge graphs, and data streams (LOS/LOF, threat & opportunity scoring, terrain/cover, trafficability, dynamic obstacles).
• Telemetry & Tooling: Add introspection/telemetry (action traces, scores, timers, counters), record-and-replay, debug overlays/heatmaps, and scenario harnesses for regression.
• Performance & Quality: Profile and optimize C++ AI systems (jobs/ECS/multithreading, memory locality, cache behavior, lock contention) to hit tight frame budgets.
• ML Inference Hooks: Integrate with model inference services, define data contracts, and gate ML policies safely behind deterministic fallbacks.
• Collaboration & Mentoring: Partner with gameplay, engine, online, and data teams; review code; mentor mid/junior engineers; uphold coding standards and test discipline

What are we looking for?

• Strong C++ (C++17/20) in large, real-time codebases; STL, templates, memory/perf, concurrency; solid debugging and profiling chops.
• Shipped features on at least one multiplayer or simulation-heavy title (AA/AAA or high-quality indie) with AI ownership.
• Deep knowledge of at least one AI architecture (GOAP, Utility AI, BTs) and practical understanding of when/why to use each.
• Experience building perception/sensing systems and scoring/utility functions, not just wiring nodes.
• Proven ability to design debuggability in from day one (telemetry, traces, visualization, reproducible test scenarios).
• Tools: Git/Perforce, CMake, Visual Studio/Clang, CI fundamentals.

What additional skills will help you stand out?

• Python 2.7 and 3.10+ for tooling or scripting; binding Python↔C++ and shipping safe/native bridges.
• Hands-on with ML for games (data collection, labeling, offline training, online inference, A/B evaluation); familiarity with RL/imitation learning concepts.
• Experience with inference runtimes (ONNX Runtime, Triton), serialization (protobuf/flatbuffers), and service plumbing (gRPC/HTTP, Kafka/Kinesis).
• Navigation for vehicles at scale (terrain analysis, cost fields, flow fields, RVO/ORCA).
• Prior end-to-end AI work on shipped or open-source titles (e.g., OpenTTD/Quake mods) showing full ownership from design to debugging.
• Exposure to ECS/job systems, lock-free structures, and SIMD; Windows/Linux cross-platform builds.

Tech You'll Touch

• C++17/20, custom engines, internal ECS/job systems
• Python 2.7 (tooling/scripting on one title), 3.10+ (machine learning)
• Build/CI: CMake, [your CI]
• Observability: internal telemetry, replay tools, debug UIs/overlays
• (Optional) ML: ONNX/TorchScript/TensorRT, gRPC/REST service

Expected Outcomes (First 12 Months)

• Ship a robust AI behavior slice (e.g., target selection + driving + firing loop) with clear KPIs (engagement quality, fairness, perf budget).
• Deliver a sensor & scoring framework that other engineers/designers can extend safely.
• Stand up AI telemetry & replay that shortens bug reproduction cycles from days to hours.
• Integrate and harden one ML-backed component behind toggles and fallbacks.

What to Include with Your Application

• Short note on an AI system you owned: problem, architecture choice, trade-offs, and how you measured success.
• Links to demos, videos, or repos (including open-source or modding AI work).
• (Optional) A brief example of a telemetry view or replay harness you built.

Work mode

• Remote (from anywhere within the selected country/EU).

Benefits

Benefits and perks are tailored to the local market and culture. Our benefits in Belgrade include:

• Additional vacation days based on years of service at Wargaming: up to 5 days on top of the statutory minimum
• Additional paid time off (5 Personal Days, Birthday Leave, Marriage Leave, Compassionate Leave)
• Sick Leave Compensation, Maternity Leave Benefits
• Premium Private Health Insurance
• Career development and education opportunities within the company
• English clubs and platform for learning languages
• Mental well-being program (iFeel)
• Commuting allowance
• Company events
• FitPass membership
• Discounts for employees
• Personal Gaming Account
• Coffee, fruits, and snacks in the office
• On-site canteen with subsidized prices for food and drinks
• Seniority Awards
• Referral program - you can recommend the best talents to the Company and receive a reward

##### Please submit your CV in English to ensure smooth processing and review.