AI Engineer
Patlytics, Inc.3 months ago
Responsibilities
- Develop and deploy scalable AI/ML algorithms for intellectual property applications.
- Design and iterate on multi-step LLM pipelines for patent claim analysis, infringement detection, and prior art search.
- Build evaluation frameworks with success metrics, golden datasets, and continuous regression testing.
- Develop hybrid retrieval systems using vector search, BM25, and re-ranking for complex patent corpora.
- Architect and ship agentic workflows involving tool use, memory, and multi-turn reasoning.
- Develop classification techniques and post-training approaches for current LLMs.
- Collaborate with patent attorneys and domain experts to create prompts, rubrics, and fine-tuning datasets.
- Own production inference cost, latency, and quality tradeoffs, including prompt caching, context management, batching, and model migration.
- Build tools to monitor data pipeline performance, data quality, and production models.
- Implement real-time and batch decisioning solutions with data engineers and the platform team.
- Perform unit testing, profiling, parameter tuning, computational experiments, and ongoing ML methodology improvements.
- Lead R&D into new technologies, data sources, data science methods, and optimization tools.
- Partner with researchers, applied scientists, engineers, product managers, and leadership on product requirements and the long-term technical roadmap.
Requirements
- Experience shipping LLM-powered features used by real customers, including production pipelines with monitoring and fallbacks.
- Strong Python skills and sufficient backend experience to own services end-to-end, including API design, data pipelines, asynchronous processing, and observability.
- Strong technical background building ML and AI pipelines, with a solid understanding of ML methods, statistics, and the ML project lifecycle.
- Experience deploying, monitoring, and maintaining data science products in cloud environments such as AWS and GCP.
- Understanding of data transformations and analytics using tools or languages such as Pandas, Scikit-learn, SQL, and Spark.
- Familiarity with database modeling, data warehousing principles, and SQL.
- Ability and willingness to develop expertise in patent and intellectual property reasoning.
- Experience with LLM model selection, prompt architecture, structured outputs, evaluation, fine-tuning or post-training, retrieval, and production inference tradeoffs.
Benefits
- Comprehensive medical, dental, and vision coverage with FSA, commuter benefits, and health advocacy through Rightway.
- Mental health and wellness support through Spring Health and Headspace, plus Mental Escape Days.
- Immediate 401(k) enrollment with no waiting period.
- Unlimited PTO, 12 paid company holidays, and a week-long Holiday Break.
- Paid parental leave.
- Professional development budget, gym membership stipend, and learning opportunities.
- Birthday and work anniversary recognition and employee referral bonuses.
- Hybrid work environment, with remote work potentially available depending on location.
Tech Stack
Categories
About Patlytics, Inc.
Patlytics builds an AI-native patent intelligence platform that supports the full IP lifecycle—prior art search, claim charting, infringement/invalidity analysis, drafting, office actions, freedom to operate, and portfolio strategy. It sells subscription software and services to corporate IP teams and law firms, delivering citation-backed outputs practitioners can audit. Founded in 2024 and headquartered in New York, the privately held company counts Quinn Emanuel, Rivian, and Canon among its customers.