The Role Own Task Management: the layer that decides what work goes to which agent, in what order, right now. Robots, automated systems and humans. Customers do not buy a robot. They buy throughput. Task Management is the decision engine behind it. Get it right and a site absorbs a peak-season order profile without anyone noticing. Get it wrong and you have idle agents on one side of the floor, a queue building on the other, and orders missing their cutoff. Navigation decides how an agent gets there. You decide where it should go and why that job beat the others. Those two areas sit next to each other and constrain each other, which makes this one of the more compelling problems in the company. What you will own
• The allocation logic as a product. Task decomposition, assignment, prioritisation, batching and rebalancing. Not just the rules, but the model of how work should flow through a site.
• Contention and fairness. Hundreds of agents, mixed capability, competing orders and hard SLA cutoffs. You define how the system decides who waits.
• The measurement system. Order cycle time, SLA attainment, agent utilisation and idle time, queue depth, starvation, priority inversion, tasks per agent hour. Define what good looks like and make it observable.
• Internal tooling. Decision replay so the team can answer "why did agent 47 get that job", order-profile simulation, and what-if analysis on a new customer's order mix before we commit to numbers.
• The interface with Navigation and Peripherals. Your assignments create their load. You will negotiate those contracts rather than throw work over a wall.
• Trade-off calls across throughput, order latency, fairness, agent wear and compute cost. What we expect in the first six months
• A baseline on utilisation and SLA attainment that leadership and the field both trust, with numbers nobody argues about.




• A simulation setup where a proposed allocation change can be evaluated against a real order profile before it reaches a customer floor.
• One measurable improvement a customer feels, in cycle time, utilisation or peak capacity.
• A clear position on where allocation should go next, and what we stop doing to get there. What you need Must have
• 7+ years in technical product management, or engineering then product, with real time on systems that control physical machines.
• You have shipped software that assigns work to physical resources. Robots, vehicles, riders, machines or crews. Something with travel time, capacity limits and a cost to being wrong.
• Fleet-scale experience. Assigning work to ten things is a different problem from assigning work to four hundred that interfere with each other.
• Depth on scheduling, task allocation, queueing, constraint solving or matching. Not writing the solver, but knowing when a greedy heuristic will stop holding and what it will cost.
• Demonstrated ability to build a measurement system from nothing. You have defined metrics, instrumented a system, and changed a team's behaviour with the data.
• Production environment exposure. You have debugged a live floor or a live network with operators standing next to you.
• Platform and API thinking. Task Management is consumed by other systems and must hold a contract.
• Written communication that is short and decisive. Strong preference
• Warehouse automation, AMR/AGV fleets, intralogistics or manufacturing execution.
• Simulation and scenario testing as a product surface, not just a QA activity.
• Heterogeneous fleets where agents have different capabilities and you did not control the hardware.
• You have run an allocation change against a live customer and know what that felt like. Nice to have
• CS, robotics, industrial engineering or EE degree.
• Operations research, optimisation or combinatorics background.
• Prior engineering role.
• Comfort with India and US

working hours
, since customers and part of the product org sit in the US.

📌 Senior Product Manager (Gurugram)
🏢 GreyOrange
📍 Gurugram