Senior AI / Software Engineer (Python, OCR, LLM & GPU Systems) (Remote)
Education:
If you did your undergrad in India, save your time and DO NOT APPLY You will have to submit a W-9 when hired, no OPT, W-8, or other types of sponsorship Senior AI / Software Engineer (Python, OCR, LLM & GPU Systems) NextGen Coding Company
Location: New York City — Hybrid / In-Person Required
Compensation: $20–$25/hour
Engagement Type: Part-Time, Approximately 20 Hours/Week Opportunity: Potential to expand hours based on performance and project needs
Work Authorization:
Candidates must be legally authorized to work in the United States under an arrangement compatible with a W-9 contractor engagement. NextGen Coding Company cannot provide OPT or employment/visa sponsorship for the role.
Role Overview:
NextGen Coding Company is hiring a highly technical AI / Software Engineer in New York City to help build a large-scale document intelligence and AI platform.
The engineer will work heavily with Python, OCR, document processing, data pipelines, LLMs, search/RAG, AWS, and NVIDIA GPU infrastructure.
A critical requirement is supporting two versions of the platform:
- On-prem / air-gapped: running local models such as Qwen on NVIDIA H100/H200 infrastructure
- Cloud: running in AWS with Claude, Qwen, OpenAI, Gemini, and other LLM/model integrations
- The underlying platform must be portable across both environments.
Responsibilities:
- Build production Python services, APIs, and data-processing pipelines
- Process large volumes of PDFs, HTML, images, scans, and structured/unstructured data
- Build OCR, extraction, parsing, normalization, and document-classification pipelines
- Extract tables, entities, relationships, metadata, citations, and structured records
- Build ETL and high-volume batch-processing systems
- Develop hybrid search, embeddings, vector search, RAG, reranking, and evidence retrieval
- Integrate Claude, Qwen, OpenAI, Gemini, and open-source models
- Deploy and serve local LLMs on NVIDIA H100/H200 GPUsWork with vLLM, PyTorch, Hugging Face, CUDA, quantization, batching, and GPU inference optimization
- Build and maintain the AWS/cloud deployment
- Work with PostgreSQL, OpenSearch, S3, Redis, queues, and related data infrastructure
- Build APIs, webhooks, Stripe integrations, authentication, and third-party integrations
- Write automated tests covering OCR, extraction, retrieval, data pipelines, and AI outputs
- Debug complex AI, data, backend, and infrastructure issues
- Document architecture and implementation decisions
Required Experience:
- Strong Python engineering experience
- Strong backend and data-engineering fundamentals
- Experience building production softwareOCR / document intelligence experience
- Experience processing PDFs, images, and large datasets
- Production experience with LLMs, RAG, embeddings, and vector search
- Experience deploying open-source models on NVIDIA GPU infrastructure
- Understanding of H100/H200-class inference environments
- Strong AWS experience
- PostgreSQL / SQLDocker and LinuxREST APIs and third-party integrations
- Ability to independently own difficult engineering problems
- Technical StackCore: Python, FastAPI, SQL, PostgreSQL, RedisAI / LLM: Qwen, Claude, OpenAI, Gemini, Hugging Face, PyTorch, vLLMOCR / Documents: PaddleOCR, Tesseract, OpenCV, PyMuPDF, Docling or similar
- Search / Data: OpenSearch/Elasticsearch, pgvector/vector databases, S3/MinIO, ETL pipelinesGPU: NVIDIA H100/H200, CUDA, vLLM, model quantization, local inference
- Cloud: AWS, Bedrock, EC2, EKS/ECS, S3, RDS, OpenSearch, SQS, IAMInfrastructure: Docker, Kubernetes, CI/CD, Linux
- Product: REST APIs, webhooks, authentication, Stripe, third-party integrations
- React, Next.js, and TypeScript experience is helpful but not the primary focus.
Ideal Candidate:
We are looking for someone who can be given thousands of pages of PDFs, scans, HTML, images, and other data plus access to an H100/H200 server and an AWS environment — and independently help engineer the system required to process, OCR, structure, search, analyze, and query the information using modern AI models. The ideal candidate is stronger in Python, AI, data processing, OCR, backend systems, and infrastructure than traditional frontend development. You should be comfortable moving between an OCR pipeline, Python data processing, Claude/AWS integration, Qwen running on an H100/H200, databases, APIs, and production infrastructure. About NextGen Coding Company NextGen Coding Company is a U.S.-based software engineering firm building custom software, AI systems, automation platforms, and enterprise applications. Projects span AI/ML, document intelligence, data engineering, financial and compliance technology, cloud infrastructure, and enterprise software. Application Requirements Applicants should provide: Resume or LinkedInGitHub and/or technical portfolio Examples of production AI/data systems built Experience with Python, OCR, LLMs, AWS, and GPU infrastructure Specific NVIDIA GPU/model deployment experience, if applicable Qualified candidates will participate in a technical discussion and engineering review before engagement.