Job Opportunities

AI Developer

Position Summary:
We are seeking an AI Engineer to design, build, and deploy offline-first, clinical AI systems running on resource-constrained edge hardware. In this role, you will bridge cloud pipelines and edge runtimes: precomputing, embedding, and indexing clinical reports and audio assets in the cloud, and delivering optimized local inference (GGUF, ONNX, Whisper.cpp) directly to disconnected field devices. You will build safe and secure low-latency retrieval systems where network access is intermittent or absent, hallucinations are unacceptable, and citation transparency is mandatory.
In this role you will collaborate with data scientists, software engineers, and product teams across various domains.

Key Responsibilities:
Edge Inference Deployment: Deploy, benchmark, and optimize local inference runtimes (llama.cpp, ONNX runtime, whisper.cpp) on low-power edge platforms (Linux, Android, low-spec x86/ARM).
Model & Data Artifact Pipelines: Automate the export of ONNX embeddings and rerankers, GGUF model quantization, and static vector store snapshot generation (ChromaDB, SQLite, or HNSW).
Hybrid RAG & Grounding: Build and evaluate hybrid retrieval pipelines (BM25 + dense vector + cross-encoder reranking) paired with strict citation mapping and deterministic safety guardrails.
Low-Bandwidth Synchronization: Design robust sync mechanisms (delta updates, compressed snapshot packaging) to push updated models and indexes from cloud pipelines to edge devices over unstable networks.
API & System Integration: Develop lightweight local APIs (FastAPI / REST / MCP) and user-facing streaming components with robust logging, debugging, and model hot-swapping.
Engineering Standards: Write maintainable, well-tested code, participate in design and code reviews, and maintain automated regression evaluation suites.

Required Skills & Experience
Local & Edge Inference: Hands-on experience running models locally via llama.cpp, ONNX runtime, or Whisper.cpp on CPU-only or memory-constrained hardware.
Advanced RAG & Vector Search: Proven experience building hybrid search pipelines (sparse + dense), cross-encoder reranking, and managing vector databases/snapshots (ChromaDB, FAISS, or SQLite-vec).
Quantization & Model Artifacts: Practical knowledge of GGUF quantization levels (K-quants), ONNX graph optimization, and bi-encoder/cross-encoder export workflows.
Clinical Guardrails & Grounding: Experience implementing structured outputs, strict citation tracking, and safety guardrails (e.g., NeMo Guardrails) to eliminate hallucinations.
Backend & Systems Python: Strong proficiency in Python (FastAPI, Pydantic, asyncio, pytest) and multi-stage containerization using Docker.
Offline/Sync Fundamentals: Understanding of differential sync, immutable snapshot delivery, or delta compression over unreliable connections.
Engineering Practices: Solid grounding in Git version control, CI/CD automation, and test-driven development.

Preferred Qualifications:
3+ years of software/AI engineering experience shipping models into real-world production environments.
Bachelor’s or Master’s in Computer Science, Software Engineering, or related technical discipline.
Familiarity with C/C++ build toolchains (CMake, Make) or compiling native libraries for target architectures (x86_64, ARM64).
Exposure to audio preprocessing pipelines (16 kHz mono PCM downsampling, VAD segmentation) or healthcare data protocols.
Experience with cloud container registries and pipeline orchestration (e.g., automated artifact generation in GitHub Actions or AWS/GCP).
Good understanding of AI practices and tools.
Experience working in an Agile development environment.
AI/ML certifications are an added advantage.

What We Offer:
A dynamic and collaborative team environment.
High-impact engineering work directly supporting community healthcare workers in remote, disconnected clinics.
Opportunities to work on innovative AI projects with real-world impact.
Professional development and learning opportunities.
Competitive salary and benefits package.

To Apply:
Submit your resume, a brief cover letter, and links to GitHub or projects showcasing your work to hr@planetmedia.in