Software Engineer- AI- First Application
- Posted
- Date not provided
- Last verified at source
- 6 hours ago
Key Responsibilities
Work as an AI- First Engineer
- Use AI coding agents and AI-driven automation as core engineering tools — to write, port, and refactor code, to draft technical documentation, and to accelerate repetitive analysis and benchmarking work.
- Apply strong technical judgment to validate AI-generated code, documentation, and analysis — knowing what to trust, what to double-check, and how to independently verify claims against raw data before they go out.
- Continuously look for opportunities to automate manual engineering workflows using AI agents and scripting, freeing up time for higher-judgment work.
Build fluency with, and help steward, Morphing's proprietary AI assistant - Develop deep, hands-on fluency with Morphing's custom AI bot, using it as a first port of call for development, analysis, and documentation work — distinct from, and in addition to, general-purpose AI coding agents.
- Take direct ownership of maintaining and incrementally improving the bot, treating it as internal infrastructure the team depends on rather than an off-the-shelf tool.
Build and optimize real application workloads - Work hands-on with code to identify application bottlenecks, performance gaps, and optimization opportunities, from application logic down to instruction- and memory-transaction-level behavior.
- Translate application behavior into concrete hardware/software requirements that hardware architects and compiler engineers can act on.
Maintain software stack and SDKs - Build front-end and back-end components wherever the platform needs them — from SDK interfaces and developer tooling to internal dashboards for workload and benchmark visualization.
- Maintain a working command of the SDKs under active development and of Morphing's broader software stack — compiler, runtime, drivers, and APIs — so that application-level work stays grounded in how the platform is actually built and used.
- Help define how the SDK and toolchain should be packaged and delivered to customers, including generating and validating the technical documentation — user guides, API references, integration notes — that ships alongside it.
Map workloads to accelerator architecture - Analyze how workloads map to modern accelerator concepts such as streaming multiprocessors, SIMT execution, warp scheduling, memory hierarchy, occupancy, arithmetic intensity, and effective utilization.
- Compare Morphing Machines' architecture against baseline GPGPU approaches for relevant workloads.
- Identify which workload characteristics make an application a strong or weak fit for the platform.
Benchmark, measure, and validate performance - Design and run benchmarks for AI, LLM inference/training, and other compute-intensive workloads, spanning simulation, emulation, and silicon where available.
- Measure latency, throughput, utilization, memory behavior, scaling behavior, and compute-bound versus memory-bound characteristics, down to the instruction and memory-transaction level where it matters.
- Produce clear technical reports explaining performance results, bottlenecks, and recommended next steps — using AI tools to speed up drafting while personally validating every number and claim before it is shared.
Support product-market-fit discovery through technical evidence - Convert customer/problem statements into testable technical hypotheses.
- Build proofs of concept, demos, and workload prototypes that demonstrate where the platform creates meaningful advantage.
- Help identify the most promising early application areas based on real performance data, implementation feasibility, and customer relevance — framed, where useful, in silicon-economics terms such as performance-per-watt and cost-per-token, with every claim backed by a verifiable proof point.
Bridge application engineering and architecture teams - Work closely with hardware architects, RTL and verification engineers, compiler/runtime engineers, and application developers to communicate workload needs.
- Provide structured feedback on architecture features, software tooling, APIs, and developer experience.
- Help the team understand what application developers will need in order to successfully adopt the platform.
Track and evaluate relevant technology trends - Maintain working knowledge of GPU and other accelerator architectures — including relevant process node, packaging, and memory technology trends — especially NVIDIA Blackwell and other platforms relevant to AI workloads.
- Evaluate emerging workload areas including AI agents, LLM serving, blockchain/crypto workloads where relevant, and other high-compute application domains.
- Separate genuinely relevant trends from hype by testing practical workload fit and technical feasibility.
Create internal technical collateral - Document hands-on findings so that engineering, leadership, and customer-facing teams can make better product and prioritization decisions.
- Contribute to demo narratives only where backed by practical implementation and measurable results.