Overview
Software Engineer (Backend) role at Parspec to design, build and scale backend systems for its AI-powered construction materials platform, driving features from estimation through production deployment.
What you'll do
- Design, develop and maintain backend systems, including APIs and messaging solutions, to support new features with modular, maintainable code.
- Implement design patterns to build high-performance systems with sub-second latency.
- Build backend systems capable of scaling efficiently to meet increasing user demands.
- Partner with the AI team to deploy, scale and optimize model inference services in production, balancing latency, throughput and cost across CPU and GPU workloads.
- Lead feature and project lifecycles from estimation to delivery in collaboration with cross-functional teams.
- Architect and optimize distributed backend systems for low latency, high availability and maintainability.
- Act as the key point of coordination between Frontend, AI and Backend teams to deliver cohesive features.
- Maintain high code quality through reviews, modular design and adherence to best practices.
- Deploy, maintain and optimize backend services and infrastructure on AWS/GCP platforms.
- Ensure thorough testing of features and tackle debugging challenges in production systems.
What you'll need
- Proficiency in backend software development using Python, Golang, or similar object-oriented languages.
- Strong understanding of RESTful APIs, service-oriented architecture (SOA) and microservice architectures.
- Hands-on experience with MySQL databases, including schema design and query optimization.
- Experience deploying and managing services on AWS/GCP platforms.
- Familiarity with containerization tools like Docker and CI/CD pipelines.
- Skilled in performance optimization and debugging within distributed systems.
- Proven ability to lead technical projects and mentor junior team members effectively.
- 3+ years of experience as a backend engineer.
Nice to have
- Exposure to AI/ML pipelines and Elasticsearch integration.
- Experience serving ML/LLM models in production using inference frameworks such as NVIDIA Triton, vLLM, TorchServe or ONNX Runtime, including request batching, autoscaling and latency/throughput tuning.
- Working knowledge of GPU hardware (e.g. NVIDIA T4, L4, A10G, A100/H100): how memory, compute and precision (FP16/INT8) trade off against cost per inference, and familiarity with optimizations such as TensorRT, quantization and GPU utilisation monitoring.
Details
- Hybrid work setup with flexible hours; preferred location Bellandur, Bengaluru, with regular in-office presence.
Read the full description and apply on the company’s own careers page.