FlytBase

- Engineering
- Pune, India (Onsite)
- Full time
- 1-2 years

What We Do : The Role :

- You'll build the brain that tells drones what to do next—not through pre-programmed routines, but through multi-agent AI systems that reason, perceive, and act autonomously. This isn't research AI or chatbot wrappers. You're building production frameworks where AI agents coordinate fleet operations, make real-time decisions based on sensor data, and execute missions across 300+ enterprise sites in 40+ countries. You'll work at the intersection of LLMs, computer vision, and robotics—designing agentic workflows that bridge reasoning models with physical drone control systems. One day you're building a perception agent that interprets thermal camera feeds to detect infrastructure faults. The next, you're orchestrating multi-agent systems where planning agents coordinate with execution agents to optimize battery life across a fleet. This is AI that flies, not AI that recommends.

What You'll Own :

- Design and build multi-agent AI frameworks that control autonomous drone operations end-to-end - Integrate LLMs, vision models, and sensor fusion systems into unified agentic architectures - Build reasoning pipelines where AI agents make mission-critical decisions in real-time with safety constraints - Develop perception modules that turn raw drone telemetry and camera feeds into actionable intelligence - Create orchestration layers that coordinate multiple AI agents—planning, execution, monitoring—into coherent workflows - Deploy and scale agentic systems across cloud and edge infrastructure with sub-second latency requirements - Build evaluation frameworks to test agent behavior, decision quality,



and safety boundaries before production - Work with robotics and platform engineers to close the loop between AI decisions and physical drone actions

Best Suited For :

- 1-2 years building production AI/ML systems—internships, side projects, or early-career roles where you shipped models to users - Strong Python engineering skills with experience in AI frameworks like LangChain, LangGraph, CrewAI, or built your own agent orchestration from scratch - Hands-on experience with LLMs (OpenAI, Anthropic, open-source models) and know how to prompt, fine-tune, and evaluate them for real tasks - Familiarity with computer vision or sensor processing—you've worked with image models, video streams, or multi-modal data pipelines - Comfort with deploying ML systems in production—you understand Docker, APIs, cloud/edge infrastructure, and real-world performance constraints - Genuine curiosity about robotics, drones, or physical AI—you want to build systems that interact with the real world, not just pixels on a screen - High learning velocity—you pick up new tools, research papers, and frameworks fast without needing a syllabus - Bonus: Experience with ROS, drone SDKs, real-time systems, or reinforcement learning

What Makes You a Flyter :

- Builder mindset: You don't wait for perfect specs. You prototype fast, test in the wild, and iterate based on what breaks. - Comfortable with ambiguity: Multi-agent AI is bleeding edge. You're excited to figure out architectures that don't have Stack Overflow answers yet. - Outcome obsession:



You measure success by whether the drone completed the mission, not whether your model hit 95% accuracy in a notebook. - AI-native thinking: You treat AI as infrastructure, not a feature. You automate your own workflows and expect AI to do the heavy lifting. - High agency: You see a gap between what AI can reason about and what drones can execute—and you build the bridge without being told how.

What We Don't Want :

- ML engineers who only work in notebooks and hand off deployment to someone else - People who treat agents as a buzzword but haven't actually built orchestration systems that coordinate multiple models - Developers who need perfect requirements and get stuck when the problem space is ambiguous - Anyone who thinks 'AI safety' means not shipping until everything is perfect—we ship fast with guardrails, not slow with paranoia - Engineers who want to work on one narrow ML problem for months—this role requires full-stack thinking from model to deployment to physical action

Why FlytBase? :

- H3 Philosophy: We're building a company that's Happy, Healthy, AND High-Performing. Not as a balance—as a harmony. We move fast without burning out. - Radical Ownership: You won't need permission to experiment, build, or lead. If you see a problem, you own the solution. - Mazaa Aayega: Our internal motto. Yes, the work is intense. But it's also deeply meaningful and genuinely fun. - Systems That Scale: We don't celebrate one-off heroics. We build compounding systems that create leverage. - In-Person Energy: We work from our Pune office because proximity accelerates everything—ideas, feedback, trust, and impact, but travel globally for customer meetings, tradeshows, and deployments.

**The Per

📌 Flink Builder (Pune)
🏢 FlytBase
📍 Pune