Eric Zheng
I’m Eric Zheng, a senior software engineer with 12 years of experience taking production systems from concept to enterprise scale, spanning ERP, professional services, retail energy, healthcare, and financial data.
About
I’m Eric Zheng, a senior software engineer with 12 years of experience taking production systems from concept to enterprise scale, spanning ERP, professional services, retail energy, healthcare, and financial data. I focus on building reliable, low-latency pipelines and regulated, high-trust systems where correctness, security, and observability matter.
Recently, I’ve been deeply focused on applied AI/ML—building LLM assistants, retrieval-augmented generation and grounding pipelines, and agentic workflows that serve tens of thousands of users. I’m comfortable across .NET, Python, and modern ML tooling, and I enjoy connecting research, engineering execution, and domain stakeholders to deliver measurable impact.
- 12 years' experience
- Joined May 2026
Industry experience
- Financial Services
- Professional Services
- Software & Internet
- Energy & Utilities
- Healthcare
- Retail
- Media & Entertainment
- Other
More experts in Miami, United States
- AI Developers for hire in Miami, United States
- AI Engineers for hire in Miami, United States
- App Developers for hire in Miami, United States
- Back-End Developers for hire in Miami, United States
- Cloud Developers for hire in Miami, United States
- Data Engineers for hire in Miami, United States
- Data Scientists for hire in Miami, United States
- DevOps Developers for hire in Miami, United States
- Full Stack Developers for hire in Miami, United States
- Software Engineers for hire in Miami, United States
Work experience· 8 roles
Anchored applied AI/ML engineering for the Bloomberg Terminal: built generative AI capabilities for news ingestion, structured-data query workflows, and real-time earnings analysis. Productionized pipelines serving 50,000+ documents per day with median end-to-end latency under 4 seconds from EDGAR publication to analyst dashboards. Re-engineered legacy extractive summarization into a hybrid extractive-then-abstractive architecture to constrain outputs to source spans, holding numerical hallucination rates below 0.5% in production across millions of summaries. Upgraded summarization shortly after the 2023 BloombergGPT release by improving summary quality scores by 28% while lowering inference cost per document. Trained a fine-tuned BloombergGPT variant to translate natural-language analyst questions into Bloomberg Query Language (BQL), reaching 25,000+ terminal users in a year and deflecting ~40% of inbound BQL support tickets. Embedded constrained decoding logic to bias token generatio
Led healthcare software modernization for an EMR/EHR and practice-management vendor serving 9,000+ specialty providers, navigating COVID-driven pivots and federal interoperability requirements. Launched Nextech’s first HIPAA-compliant telehealth platform in 7 weeks during lockdown, onboarding 600+ specialty practices in the first 90 days and exceeding 100,000 video visits in 6 months. Designed a decoupled sidecar architecture integrating with legacy monolithic EMR via REST APIs and webhooks to avoid direct coupling and reduce refactoring effort while preserving visit traceability in provider charts. Composed video stack using Twilio Programmable Video with .NET and React session handling while keeping PHI within EMR boundaries, improving security posture and shortening the typical security review cycle. Headed FHIR R4 API implementation ahead of 21st Century Cures Act deadlines, supporting 50,000+ FHIR API calls per day from third-party apps. Devised a translation layer mapping propr
Led healthcare software modernization for a Tampa-based EMR/EHR and practice-management vendor serving 9,000+ specialty providers. Launched NextTech's first HIPAA-compliant telehealth platform in 7 weeks during COVID lockdown, onboarding 600+ specialty practices in the first 90 days and surpassing 100,000 video visits in 6 months as a permanent product line. Designed sidecar service architecture to integrate with legacy monolithic EMR via REST APIs and webhooks (avoiding direct schema coupling) while keeping video visits indistinguishable from in-person visits in provider charts. Composed the video stack using Twilio Programmable Video, .NET Core 3.1, and a React+session model approach that kept PHI within EMR boundaries and reduced typical 4-week security review time to under 1 week. Headed FHIR R4 API implementation ahead of the 21st Century Cures Act deadline, preserving NextTech's ONC certification and serving 50,000+ FHIR API calls per day from third-party patient apps within 6 mo
Education
Hire Eric Zheng today
Post a job and invite Eric, or contact them directly. No hidden commission.









