Machine Learning Engineer

Piper Companies is currently looking for an Machine Learning Engineer in Washington DC to design, build, and operationalize scalable AI/ML solutions across a variety of mission-critical applications. This is a hybrid role and candidates will undergo a federal background check to receive a security clearance. Candidates with a prior or active clearance are preferred.

Responsibilities for the Machine Learning Engineer:

• Collaborate with data scientists and subject matter experts to design, build, and train machine learning systems
• Implement and optimize Large Language Models (LLMs), Retrieval-Augmented Generation (RAG) systems, and AI agent architectures for enterprise use cases
• Deploy ML solutions via MLflow, AWS SageMaker, or custom APIs
• Document ML artifacts, processes, and performance outcomes clearly and comprehensively

Qualifications for the Machine Learning Engineer:

• 5+ years of experience in ML Engineering or Applied Machine Learning and Python programming
• Hands-on experience with ML frameworks (e.g., scikit-learn, XGBoost, PyTorch, TensorFlow)
• Hands-on experience with training frameworks such as TensorFlow, PyTorch, or Hugging Face
• Proficiency with Databricks, MLflow, and PySpark
• Practical experience building and deploying LLMs, RAGs, and AI agent systems
• Experience with AWS services such as S3, EC2, Lambda, SageMaker, and Step Functions for scalable ML workloads

Compensation for the Machine Learning Engineer:

• Salary Range: $150,000-170,000 (depending on experience)
• Comprehensive benefit package; Cigna Medical, Cigna Dental, Vision, 401k w/ ADP, PTO, paid holidays, sick Leave as required by law

This job opens for applications on 2/13/26. Applications for this job will be accepted for at least 30 days from the posting date

#LI-BM2

#LI-HYBRID

transformers, attention mechanisms, self-attention, multi-head attention, positional encoding, tokenization, subword tokenization, Byte Pair Encoding (BPE), sentencepiece, embeddings, contextual embeddings, vector representations, model fine-tuning, instruction tuning, supervised fine-tuning (SFT), reinforcement learning from human feedback (RLHF), direct preference optimization (DPO), quantization, QLoRA, LoRA adapters, parameter-efficient fine-tuning (PEFT), model distillation, hallucination mitigation, prompt engineering, prompt templates, chain-of-thought prompting, zero-shot inference, few-shot learning, model latency optimization, inference serving, model sharding, distributed training, gradient checkpointing, mixed precision training, FP16, BF16, tensor parallelism, pipeline parallelism, model parallelism, GPU acceleration, CUDA kernels, vLLM, Hugging Face Transformers, tokenizer parallelization, vector databases, dense retrieval, semantic search, ANN search, FAISS, Milvus, Pinecone, Weaviate, ChromaDB, hybrid search, BM25 retrieval, chunking strategies, document splitting, embedding generation, retriever-ranker architecture, context window management, retrieval latency, indexing pipelines, retrieval evaluators, grounding, context injection, query expansion, document reranking, relevance scoring, metadata filtering, context compression, synthetic data generation for retrieval, RAG evaluation metrics, retrieval latency optimization, MLflow Tracking, MLflow Models, MLflow Registry, MLflow Projects, experiment tracking, model lineage, model versioning, artifact storage, run metadata, autologging, reproducible ML pipelines, conda environments, model deployment, model packaging, MLflow scoring server, parameters & metrics tracking, model governance, reproducibility, A/B model comparison, deployment to REST endpoints, MLflow + Databricks integration, ageMaker Studio, SageMaker Notebook Instances, SageMaker Training Jobs, SageMaker Inference Endpoints, SageMaker Serverless Inference, SageMaker JumpStart, SageMaker Pipelines, SageMaker Feature Store, SageMaker Model Registry, SageMaker Experiments, SageMaker Autopilot, model monitoring, Data Wrangler, processing jobs, distributed training on SageMaker, multi-model endpoints, batch transform, SageMaker SDK, ECR containers, IAM roles for ML, VPC-configured training, spot training, Hugging Face DLCs, SageMaker Clarify, SageMaker Debugger, Databricks Runtime, Delta Lake, Delta Tables, Unity Catalog, Databricks Workflows, Databricks Jobs, Databricks Model Serving, Databricks Feature Store, MLflow on Databricks, Databricks Notebooks, DBFS, Photon engine, Databricks SQL, vector search in Databricks, Mosaic AI, MosaicML training, distributed compute clusters, autoscaling clusters, Lakehouse architecture, Databricks Marketplace, data ingestion pipelines, clustering policies, secret scopes, job clusters, ML runtime versions