15 Nov
|
Spatial Alphabet
|
Kondapur
15 Nov
Spatial Alphabet
Kondapur
Role Summary: We are converting our drone video and image analytics MVP into a multi-tenant SaaS platform, with GA.The Python Developer (AI/ML) builds the production services around our models: GPU-based inference, media processing pipelines, and the APIs that let tenants submit drone imagery and video and receive reliable results. You will turn research-grade scripts from the MVP into scalable, observable, cost- productive services.
Key Responsibilities
1. Inference services: Build and maintain Python services (FastAPI) that serve computer-vision models (detection, segmentation, classification, and tracking) on GPU infrastructure, with batching, queueing,and autoscaling.)
2. Model packaging and serving: Containerize models and serve them with Triton, TorchServe, or ONNX Runtime/TensorRT. Support model versioning, A/B rollout, and rollback.
3. Media pipelines: Implement ingestion-to-inference workflows for large images, orthomosaics, and video, including frame extraction, tiling, geotagging with EXIF/XMP and telemetry, post-processing, and result aggregation. Use FFmpeg, OpenCV, and GDAL/rasterio.
4. Async processing: Build job orchestration on queues and workers (Celery, SQS/Kafka, or Ray) with retries, idempotency, and dead-letter handling.
5. Multi-tenant ML: Enforce tenant isolation in data access, model usage, and results. Support per- tenant quotas, usage-metering hooks (inference minutes and GPU seconds), and per-tenant model configuration.
6. Performance and cost: Profile and optimize latency, throughput, and GPU utilization using mixed precision, quantization,
and batching. Track cost per inference.
7. MLOps: Contribute to training-to-production pipelines: dataset versioning, experiment tracking (MLflow or W&B;), model registry, automated evaluation, and drift or quality monitoring.
8. Quality: Write unit, integration, and load tests. Validate model outputs against agreed accuracy and confidence thresholds with QA and data scientists.
9. Security and compliance: Follow secure coding practices, secrets management, encryption in transit and at rest, and audit logging in support of SOC 2.
10. Migration: Port MVP inference code and models to the new platform, and verify result parity for existing customers.
11. Collaboration: Work with the Drone Data Architect, Tech Architect, full-stack, DevOps/SRE, and Business Analyst on API contracts, result schemas, and SLAs. Participate in code reviews and sprint ceremonies.
Required Qualifications Bachelor’s or Master’s degree in Computer Science, Data Science, Engineering, or a related field.
4+ years of professional Python development; strong in typing, async I/O, packaging, and testing (pytest).
2+ years deploying ML or computer-vision models to production (PyTorch or TensorFlow).
Experience with FastAPI or Flask, REST/gRPC APIs, and PostgreSQL.
Hands-on experience with Docker and Kubernetes; GPU workloads (CUDA basics, NVIDIA drivers/runtime).
Experience with OpenCV, NumPy, and video/image processing at scale.
Cloud experience on AWS or Azure (S3/Blob, queues, container services, GPU instances).
Git, CI/CD, and code-review discipline.
Preferred Qualifications:
Detection and segmentation frameworks (YOLO, Detectron2, MMDetection, SAM) and object tracking.
Model optimization with ONNX, TensorRT, or quantization.
Geospatial tooling (GDAL, rasterio, GeoPandas, PostGIS) and drone data formats.
Distributed processing with Ray, Spark, or Dask.
Monitoring with Prometheus, Grafana, or OpenTelemetry.
Experience with multi-tenant SaaS and usage-based billing integrations.
Infrastructure as code (Terraform) and MLflow or Kubeflow.
Utility, telecom, or infrastructure inspection domain experience.
Key Deliverables:
Production-ready GPU inference services and model-serving components.
Reliable media ingestion, processing, and asynchronous job workflows.
Tenant-isolated ML execution with usage-metering hooks and model configuration.
Observable, tested, and cost-aware services that support migration from the MVP to the GA platform.
Education: Bachelor’s or Master’s degree in Computer Science, Data Science, Engineering, or a related field.
Pay: ₹800,000.00 - ₹1,200,000.00 per year
Work Location: In person
📌 Python Developer (AI/ML) (Kondapur)
🏢 Spatial Alphabet
📍 Kondapur