Back End Developer

About Sharedpro

Sharedpro is an AI platform company building a suite of AI-native products on a shared foundation of computer-vision and audio-ML infrastructure. We build a family of products that sit on the same hard backend problems: real-time inference, vector search, GPU model serving etc

Our products include:

• AutoPilot Interview an AI-powered interview room with real-time voice AI, live proctoring, and face detection.
• Emotions API a 48-emotion recognition engine model.
• TrueFace face verification.
• Emotion Quest a gamified, multiplayer emotion-mimicking experience.
• Pediatric Developmental screening early ASD / milestone tracking from motion, audio, and gaze signals.

We're a small, fast-moving team where engineers own real systems end to end and ship every week.

The Role

We're looking for a Backend Developer comfortable across a Python stack and excited to work close to the metal on real-time, ML-powered systems. You'll build and scale the services that power our products from REST and WebSocket APIs to model-serving pipelines and vector search.

What You'll Do

• Design, build, and scale backend services in Python and related frameworks (e.g., FastAPI, Flask, Django).
• Integrate and optimize ML model serving on our in-house GPU servers NVIDIA Triton Inference Server, ONNX Runtime, and TensorRT with dynamic batching and FP16 optimization
• Develop vector search and RAG systems with Qdrant, including hybrid dense + sparse retrieval and rerankers
• Design clean REST APIs, data models, and session / state management
• Work with PostgreSQL and MySQL schema design, migrations, and query performance
• Containerize services with Docker and maintain Jenkins CI/CD pipelines
• Handle concurrency the right way virtual threads, async patterns, and safe concurrent inference
• MLOps pipeline data collection, training, evaluation, and deployment.

What We're Looking For

Must-have

• Working proficiency in Python (FastAPI a strong plus)
• Strong grasp of REST API design and HTTP fundamentals
• Hands-on with SQL databases (PostgreSQL or MySQL)
• Comfortable with Docker and Git
• Good understanding of concurrency, threading, and async patterns

Nice-to-have

• ML model serving Triton, ONNX Runtime, TensorRT, or DJL
• Vector databases (Qdrant or similar) and embeddings / semantic search / RAG
• WebSocket and other real-time systems
• Exposure to computer-vision or audio-ML workloads
• Familiarity with GPU inference, model formats, or throughput / batching optimization

Work Location

Vadodara, Gujarat (Work From Office)

Why Sharedpro

• In-house GPU hardware. Dedicated GPU servers for training, inference, and experimentation you're not rationing cloud quota to run a model.
• Wide ownership. Small team, big surface area. You own services end to end and watch your work ship.
• Genuinely hard problems. Multi-model inference pipelines, sub-second latency targets, hybrid vector search, document forensics the kind of work that makes you a sharper engineer.
• Modern stack, pragmatic choices. We pick the right tool for the job and aren't afraid of the bleeding edge when it earns its place.