AI Developer (3–5 YOE) | Location Pune

About the Role

We’re looking for a mid-level

AI Developer

to build, train, and deploy ML / DL models—especially across

Computer Vision

and core

Python

stacks. You’ll own model lifecycle end-to-end : problem framing, data pipelines, training, optimization, and production deployment.

Key Responsibilities

Translate business problems into ML use cases; define success metrics and baselines.

Design and implement

ML / DL pipelines

(data prep, feature engineering, model training, evaluation, and monitoring).

Build

Computer Vision

solutions (classification, detection, segmentation, OCR, tracking, quality inspection).

Develop robust

Python

services, APIs, and inference microservices; integrate with existing systems.

Optimize models for speed, accuracy, and cost (quantization, pruning, batching, ONNX / TensorRT, GPU utilization).

Create and maintain

MLOps

workflows (versioning, CI / CD, experiment tracking, automated retraining).

Document experiments, communicate results, and present insights to technical & non-technical stakeholders.

Ensure data security, model governance, and compliance with internal standards.

Must-Have Skills

Python

(NumPy, Pandas, PyTorch and / or TensorFlow / Keras).

Solid grasp of

ML fundamentals

(supervised / unsupervised learning, evaluation metrics, cross-validation, bias / variance).

Deep Learning

for CV : CNNs, transfer learning, pre-trained backbones (ResNet / EfficientNet / ViT), augmentation.

Computer Vision

toolchain : OpenCV,

object detection

(YOLO / Detectron / SSD),

segmentation

(U-Net / Mask R-CNN).

Experience packaging inference as

REST / gRPC

services (FastAPI / Flask) and deploying on

Docker

(basic Kubernetes a plus).

Hands-on with

Git , experiment tracking (Weights & Biases / MLflow), and model / version management.

Good-to-Have

OCR (Tesseract, EasyOCR), document AI, key-value extraction.

Edge / embedded deployment (NVIDIA Jetson, TensorRT, OpenVINO).

MLOps : Airflow / Prefect, Kubeflow / SageMaker, feature stores, data versioning (DVC).

NLP basics (transformers) and vector databases.

Cloud exposure (AWS / GCP / Azure) — GPUs, storage, IAM basics.

Basic frontend / backoffice UI (Streamlit / Gradio) for demos.

Qualifications

Bachelor’s / Master’s in CS, ECE, Data / AI, or equivalent practical experience.

3–5 years in applied ML / AI with at least

2+

successful production deployments.