AI Product Engineer (Visual Intelligence and Automation)

What You’ll Do

● Build and own AI-backed features end to end, from ideation to production — including

layout logic, smart cropping, visual enhancement, out-painting and GenAI workflows for

background fills

● Design scalable APIs that wrap vision models like BiRefNet, YOLOv8, Grounding

DINO, SAM, CLIP, ControlNet, etc., into batch and real-time pipelines.

● Write production-grade Python code to manipulate and transform image data using

NumPy, OpenCV (cv2), PIL, and PyTorch.

● Handle pixel-level transformations — from custom masks and color space conversions

to geometric warps and contour ops — with speed and precision.

● Integrate your models into our production web app (AWS based Python/Java backend)

and optimize them for latency, memory, and throughput

● Frame problems when specs are vague — you’ll help define what “good” looks like, and

then build it

● Collaborate with product, UX, and other engineers without relying on formal handoffs

— you own your domain

What You’ll Need

● 2–3 years of hands-on experience with vision and image generation models such as YOLO,

Grounding DINO, SAM, CLIP, Stable Diffusion, VITON, or TryOnGAN — including

experience with inpainting and outpainting workflows using Stable Diffusion pipelines

(e.g., Diffusers, InvokeAI, or custom-built solutions)

● Strong hands-on knowledge of NumPy, OpenCV, PIL, PyTorch, and image

visualization/debugging techniques.

● 1–2 years of experience working with popular LLM APIs such as OpenAI, Anthropic,

Gemini and how to compose multi-modal pipelines

● Solid grasp of production model integration — model loading, GPU/CPU optimization,

async inference, caching, and batch processing.

● Experience solving real-world visual problems like object detection, segmentation,

composition, or enhancement.

● Ability to debug and diagnose visual output errors — e.g., weird segmentation artifacts,

off-center crops, broken masks.

● Deep understanding of image processing in Python: array slicing, color formats,

augmentation, geometric transforms, contour detection, etc.

● Experience building and deploying FastAPI services and containerizing them with

Docker for AWS-based infra (ECS, EC2/GPU, Lambda).

● Solid grasp of production model integration — model loading, GPU/CPU optimization,

async inference, caching, and batch processing.

● A customer-centric approach — you think about how your work affects end users and

product experience, not just model performance

● A quest for high-quality deliverables — you write clean, tested code and debug edge

cases until they’re truly fixed

● The ability to frame problems from scratch and work without strict handoffs — you

build from a goal, not a ticket