UAV Communication Engineer — Native C++ WebRTC for Peer-to-Peer Video & MAVLink

About Manastu:

Manastu Space is a Space Safety and Logistics company focused on making space secure, sustainable, and accessible. We develop technologies to address critical challenges such as space debris mitigation and green propulsion, enabling safer satellite operations, and long-term sustainability of the space ecosystem. Our work sits at the intersection of advanced engineering, applied research, and mission-critical systems, requiring deep technical capability and long-term thinking.

- Website: https://www.manastuspace.com
- LinkedIn: https://www.linkedin.com/company/manastu-space/
- Email: [email protected]

About the role

You'll design and own the peer-to-peer network that carries live video and MAVLink between our drones and the office. Drones fly on 5G; operators watch and control them from a native desktop app, and several receivers may watch one drone at once. Latency is the number that matters most.

We don't add server hardware to this system, so you'll make peer-to-peer work across carrier NAT, using our existing cloud platform only for signaling and a relay fallback. This is a hands-on individual-contributor role: you'll build both ends, from the drone's companion computer to the receiving app.

What you'll do

- Design the end-to-end network: how drones on 5G behind carrier NAT reach office receivers, how connections are set up and recovered, and what happens when a link degrades or drops.
- Build peer-to-peer WebRTC on both ends: the drone side on embedded Linux (Jetson-class companion computers) and the receiving side in our native C++ desktop app.




- Carry MAVLink over WebRTC data channels: unordered, low-latency delivery for high-rate telemetry and reliable delivery for commands, with routing for several drones.
- Own the latency budget from camera to screen: encoder settings, packetization, jitter buffer and playout delay, congestion control and error recovery.
- Let several receivers watch the same drone without multiplying its 5G uplink, for example by having the main office receiver re-share video over the local network.
- Set up signaling and a TURN relay fallback through our cloud platform, without adding servers on site.
- Secure every link: per-device identity, encrypted media and data, and signed MAVLink commands.
- Build tools to see the network: per-link latency, loss, jitter and bitrate monitoring, and test rigs that simulate bad links.
- Work with the desktop app engineer so decoded video and telemetry arrive cleanly in the app's interface.
- Test on the bench and in field trials over real 4G/5G networks, and document the key design decisions.

What you'll need

- Bachelor's OR Master's degree in computer science, electronics or a related field.
- Minimum 3 years building and 2 years in shipping production WebRTC systems with real-time video.
- Native WebRTC in C++ using libwebrtc, GStreamer's webrtcbin or libdatachannel.




- Real-time media transport: RTP/RTCP, retransmission (NACK), keyframe requests (PLI/FIR), congestion control (GCC/TWCC), forward error correction and jitter buffers.
- Low-latency video: H.264/H.265 hardware encoding and decoding, and tuning keyframe interval, bitrate and encoder settings for latency and loss recovery.
- WebRTC data channels, including when to use reliable, ordered delivery and when to use unordered, unreliable delivery.
- MAVLink (messages, system IDs, MAVLink 2 signing), or a comparable telemetry and command protocol you've carried over data channels.
- Linux networking: UDP/TCP internals, Wireshark and tcpdump, simulating bad links with tc/netem, and debugging on embedded Linux.
- Strong, multithreaded C++.
- WebRTC used only through browser JavaScript APIs or hosted services (Twilio, Agora, LiveKit Cloud) doesn't meet this role.

Good to have

- Drone or robotics experience: ArduPilot or PX4, mavlink-router, companion computers such as Jetson or Raspberry Pi.
- Shipped real-time video over cellular: teleoperation, drones, body cameras or vehicle video.
- GStreamer pipelines on Jetson with hardware encoding (NVENC).
- Running TURN servers (coturn) and designing signaling over WebSocket or MQTT.
- SRT or RTSP, and when to choose them over WebRTC.
- Simulcast or SVC for receivers on links of different quality.
- How 4G/5G networks behave: carrier NAT, cell handovers, uplink limits and private APNs.
- Qt/QML, to help bring video into the desktop app.
- Defense or aerospace practices: approved encryption, secure boot and documented design reviews.

📌 UAV Communication Engineer — Native C++ WebRTC for Peer-to-Peer Video & MAVLink (Navi Mumbai)
🏢 Manastu Space
📍 Navi Mumbai