AIML Engineer

Magnasoft; Bengaluru; Hybrid; INR 1,800,000 - 2,600,000; Full time

Job summary

Magnasoft Bengaluru is seeking a mid-level AI/ML Engineer to build production-grade models (object detection, segmentation, OCR) and the surrounding AI backend. The role blends model work with data pipelines, emphasizes real-world reliability over notebooks, and requires strong Python skills.

You will deploy with Docker and Kubernetes on AWS EKS, design data schemas in MongoDB, and collaborate with senior engineers to improve data quality and evaluation harnesses across releases.

Qualifications

• 3-5 years hands-on building production ML/AI; you shipped models that real users rely on, not only POCs.
• Strong Python for both model and product code; backend and pipeline logic around models.
• Strong PyTorch (or TensorFlow) and solid ML fundamentals, with full lifecycle: data prep, training, evaluation, deployment.
• MongoDB: design document schemas and write non-trivial aggregation queries.
• PostgreSQL working knowledge; productive with room to deepen.
• Docker and Kubernetes (AWS EKS), and hands-on AWS; you ship and run your own code.
• Genuinely hands-on and eager to grow under a Principal AI Engineer and Senior Applied ML Engineer.
• Stack: Python, PyTorch, MongoDB, PostgreSQL, Docker/Kubernetes on AWS EKS; depth in ML and Python backend/data layer matters most.
• Strong plus: Computer Vision (detection/segmentation YOLO, Detectron2, Mask R-CNN) or OCR/document AI.
• Geospatial / GIS exposure (imagery, GDAL/geopandas, remote sensing).
• MLOps depth: MLflow, model registry, monitoring, data/label versioning.
• RAG / GenAI / agentic exposure, or data-centric ML (annotation tooling, active learning).
• Fluency with AI-assisted coding (e.g., Claude Code, Copilot, Cursor) to move faster.

Responsibilities

• Build and ship production models: object detection, segmentation, OCR/text extraction, and classification models used by products.
• Build the AI backend the models run in; run models on incoming data; write post-processing and pipeline logic in Python.
• Work in the data layer; design document structures and write queries and aggregations your pipeline and retraining loop depend on.
• Feed the data flywheel: the annotation correction retraining loop that makes the models better release over release.
• Own evaluation for your work: benchmarks, error analysis, and quality metrics tied to real product outcomes.
• Deploy and run your models and your pipeline code using Docker, Kubernetes on AWS EKS.
• Collaborate under the Principal AI Engineer's direction, and partner with the Senior Applied ML Engineer on data quality and the eval harness.

Skills

Python

PyTorch

TensorFlow

Model deployment

Data processing

Tools

MongoDB

PostgreSQL

Docker

Kubernetes

AWS

MLflow

OCR

Geospatial tools (GDAL/geopandas)

AI/ML pipelines

Job description

"

Job Summary

AI/ML Engineer (Mid) at Magnasoft Bengaluru Hybrid (WFO - 3 Days). Hands-on model building and the AI backend + pipelines.

Responsibilities

• Build and ship production models: object detection, segmentation, OCR/text extraction, and classification models behind our products. Not notebooks that die in a repo: models real customers depend on.


• Build the AI backend the models live in. Run the models on incoming data, then write the post-processing and pipeline logic that turns raw model output into clean, structured product data. All in Python.


• Work in the data layer. Detected and human-corrected results are stored in a document store (MongoDB). You design document structures and write the queries and aggregations your pipeline and the retraining loop depend on.


• Feed the data flywheel: the annotation correction retraining loop that makes the models better release over release.


• Own evaluation for your work: benchmarks, error analysis, and quality metrics tied to real product outcomes (cost-of-error, reviewer effort saved), not just headline accuracy.


• Deploy and run your models and your pipeline code using Docker, Kubernetes on AWS EKS, and iterate on what production tells you.


• Work under the Principal AI Engineer's technical direction, and partner with the Senior Applied ML Engineer on data quality and the eval harness.

Requirements

• 3-5 years hands‑on building production ML/AI; you shipped models that real users or customers rely on, not only POCs or coursework.


• Strong Python for both model and product code. You write the backend and pipeline logic around your models - post-processing, data structures, pipeline stages, APIs - not just training scripts.


• Strong PyTorch (or TensorFlow) and solid ML fundamentals, with the full lifecycle in your own hands: data preparation training evaluation deployment.


• MongoDB: comfortable you can design document schemas and write non-trivial aggregation queries.


• PostgreSQL working knowledge; comfortable enough to be productive, with room to deepen on the job.


• Docker and Kubernetes (we run AWS EKS), and hands‑on AWS; you ship and run your own code, you don't hand it to someone else to deploy.


• Genuinely hands‑on and eager to grow - you ramp fast under a strong Principal and take on more over time.


• Our stack: Python across the board; PyTorch; MongoDB; PostgreSQL; Docker/Kubernetes on AWS EKS; AWS for cloud and GPU-backed training/inference. Depth in ML and the Python backend/data layer matters most; we expect on‑the‑job growth on the rest.


• Strong plus: Computer Vision (detection/segmentation YOLO, Detectron2, Mask R-CNN) or OCR/document AI.


• Geospatial / GIS exposure (imagery, GDAL/geopandas, remote sensing).


• MLOps depth: MLflow, model registry, monitoring, data/label versioning.


• RAG / GenAI / agentic exposure, or data-centric ML (annotation tooling, active learning).


• Fluency with AI-assisted coding (e.g., Claude Code, Copilot, Cursor) to move faster.


• You might not be a fit if:
  
  
  • You only train models and hand them off. This role writes the product/backend/pipeline code the models run inside, and works daily in the data layer.
  
  
  • Your background is mostly analytics/BI/dashboards rather than building and shipping models.
  
  
  • Your ML is purely academic or POC with nothing in production.
  
  

"