Vetted BERT Professionals

“Backend engineer who built and evolved Python/FastAPI services (including AWS-deployed ML prediction APIs) for real-time profitability and risk insights at TenXengage. Emphasizes pragmatic architecture, strong validation/observability, and secure access controls (RBAC + row-level filtering), and has led safe migrations via parallel runs and incremental rollouts; reports ~20% forecasting accuracy improvement.”

“Built and deployed an LLM-powered clinical decision support and risk monitoring platform for mental health at Valuai.io, emphasizing low-latency, evidence-grounded responses and crisis-safe behavior with clinician escalation. Strong production agent-orchestration background (LangChain/CrewAI) plus rigorous evaluation (clinician-in-the-loop + evaluator agent) and large-scale synthetic testing; also applied multi-agent workflows to document verification and fraud detection during an AI internship at Nixacom.”

“AI/back-end engineer at Canon who helped build and operate an internal production LLM platform that acts as a secure middle layer between users and models, defending against jailbreaks/prompt injection while enabling RAG, memory, and grounded responses over company data. Experienced with LangChain/LangGraph orchestration, vector DB retrieval, and reliability practices (testing, monitoring, adversarial prompts) to run high-throughput, low-latency AI workflows in production.”

“ML/NLP practitioner who has delivered production systems in regulated domains, including a healthcare compliance pipeline using RAG (GPT-4/Claude) plus TF-IDF retrieval that increased document review throughput 4.5x. Also has hands-on experience improving fraud detection data quality via entity resolution (Levenshtein, Dedupe.py) validated with A/B testing, and building scalable, monitored workflows with Airflow, CI/CD, and AWS SageMaker.”

“Built and deployed a production LLM-powered text-to-SQL/document intelligence chatbot on AWS that lets non-technical business users query complex enterprise databases in plain English. Demonstrates deep practical expertise in schema-aware prompting, embeddings-based schema retrieval, SQL safety/validation guardrails, and rigorous offline/online evaluation with human-in-the-loop approvals for risky queries.”

“Currently at Uplift AI shipping production LLM features that generate personalized growth insights from user reflections using BERT + embeddings + RAG, with strong safety/guardrail practices for sensitive contexts. Also built an end-to-end React Native UGC challenge submission/moderation system that improved repeat submissions and 7-day retention, and has applied rigorous clinical-style evaluation methods on a dental X-ray disease detection project to reduce false negatives.”

“Applied LLMs to classify long nonprofit mission statements into 8 segments without labeled data, using an ensemble of clustering/embedding methods plus zero-shot RoBERTa/BART and a Tree-of-Thought prompting pipeline with LLM-as-judge evaluation (Gemma). Also built LangChain/LlamaIndex agentic RAG workflows including a text-to-SQL data analysis assistant grounded on DB schema with retries and performance optimizations on an HPC cluster.”

“Built a production-grade, semi-automated document recognition and classification system for large volumes of scanned PDFs, starting from little/no labeled data and handling highly variable scan quality. Deployed on AWS using SageMaker + Docker and orchestrated on EKS with a microservices design that scales CPU-heavy OCR separately from GPU inference, with strong reliability controls (validation, fallbacks, retries, readiness probes).”

“Software engineer who deployed and stabilized a real-time analytics platform at Senecio Software, focusing on production reliability, observability, and performance under load. Experienced debugging issues spanning distributed services and networking (e.g., tracing timeouts to packet loss from misconfiguration) and extending Python (FastAPI/Django) APIs for customer-specific analytics features in a configurable, maintainable way.”

“LLM/agentic AI builder who deployed a production ITSM automation agent on Google ADK integrating ServiceNow and FreshService, with strong safety guardrails (human-approval gating and runbook-only command execution) and rigorous evaluation (500 synthetic tickets; 80%+ false-positive reduction). Also partnered with finance to deliver an AI agent that automated invoice/SOW retrieval and monthly reporting to account managers, reducing manual back-and-forth.”

“Built and deployed a production-grade, multi-agent Text-to-SQL assistant that lets non-technical stakeholders query large enterprise databases in natural language. Uses Pinecone-based schema retrieval + LLM reasoning (Gemini/Claude/GPT) with a dedicated validation agent (schema/syntax checks and safe dry runs) to reduce hallucinations and improve reliability, while optimizing latency and cost via async execution and embedding caching.”

“LLM engineer/data analyst who built a production RAG QA assistant over the Jurafsky & Martin NLP textbook to reduce hallucinations and provide explainable, source-grounded answers. Experienced with LangChain/LangGraph orchestration, retrieval optimization (embeddings, vector DBs, caching), and rigorous evaluation/monitoring (Retrieval@K, A/B tests, telemetry/drift). Previously communicated analytics insights to non-technical stakeholders at GS Analytics using Power BI and simplified reporting.”

“Data scientist who has built production LLM agents (GPT-4o + LangChain + RAG) to automate analyst-style ad hoc CSV analysis with guardrails and GPT-as-a-judge evaluation. Also delivered an explainable healthcare NLP system for ICD code classification by collaborating closely with clinicians, using a hybrid rule-based decision tree + BERT model to reach 97% accuracy and cut manual review time.”

“Built and deployed a production RAG-based multilingual analytics assistant for healthcare operations, enabling non-technical teams to query claims/EHR and risk metrics with grounded explanations. Demonstrates strong end-to-end LLM system engineering (retrieval tuning, re-ranking, hallucination controls, verification layers) plus workflow orchestration (Airflow/Composer/Step Functions) and stakeholder-driven iteration via prototypes and dashboards.”

“Currently at ProShare and reports building an AI/LLM-powered system deployed to production, aimed at helping with status-related difficulties and reducing misunderstandings across transactions. Also cites prior collaboration at Porsche with marketing teams, focusing on translating marketing goals into technical requirements and communicating solutions clearly to non-technical stakeholders.”

“Built and deployed a production generative-AI copilot at Tungsten that automates invoice/form extraction template creation, reducing weeks of manual model-building work. Combines fine-tuned LLMs (PyTorch/HuggingFace) with OpenCV layout grounding to reduce hallucinations, and runs an end-to-end Kubeflow-based MLOps pipeline with drift monitoring, canary releases, and automated retraining.”

“Senior backend/data engineer who has built and maintained HIPAA-compliant, real-time clinical FastAPI services on AWS, orchestrating ML/LLM and vector DB calls with strong reliability patterns (auth, timeouts/retries, graceful degradation, idempotency). Also delivered AWS IaC/CI-CD (Terraform/Helm/GitHub Actions) across EKS/Lambda/SageMaker and built Glue/Spark ETL with schema evolution and data quality controls, plus demonstrated large SQL performance wins (15 min to <9 sec) and hands-on incident ownership.”

“Built and shipped "OnCall Copilot," a production Slack-based RAG assistant that answers on-call questions from runbooks and postmortems with citations using a FAISS vector index. Emphasizes reliability and measurable performance via strict guardrails ("no evidence, no answer"), evaluation metrics, drift monitoring, and operational hardening with Docker, logging, health checks, and offline fallback.”

“Graduate-school lab engineer who built and owned the final architecture of a Microservices Hub that integrates REST APIs, issues API keys, monitors 10+ Linux servers, and visualizes service dependencies via a topology graph. Strong in bridging legacy and modern stacks (Dockerized and non-Dockerized services like Apache/screen) using deep Linux/networking knowledge, plus practical real-time audio streaming for STT/TTS and experience mentoring others.”

“Robotics software engineer who built a ROS 2 Humble autonomous hospital-equipment detection/localization robot end-to-end in Gazebo (custom worlds/models, Nav2 waypoint navigation, YOLOv8n perception, TF2-based depth fusion) and solved real-time integration issues via multithreading and QoS tuning. Also implemented and tuned an MPPI controller to enable smooth reverse parking on an OpenPodCarV2 platform, including real-world reverse engineering and hardware/software debugging.”

“Built and owned a customer-facing Document Intelligence Service for legal contract analytics at Noasis Digital, delivering extraction/summarization with careful accuracy controls (confidence thresholds, versioned deployments, production logging). Also developed a React/TypeScript document review app and internal QA dashboard, and has hands-on microservices experience with async messaging (RabbitMQ), timeout tuning, and centralized structured logging for reliability at scale.”

“AI/LLM engineer who built and deployed a production healthcare RAG chatbot ("DoctorBot") with strict medical safety guardrails, an 85% confidence-gated verification layer, and latency optimizations that cut responses from ~8s to ~2–3s. Also worked on finflow.ai to generate finance/banking test cases from BRDs, collaborating closely with non-technical domain stakeholders, and has hands-on orchestration experience with LangChain/LangGraph and agentic evaluation/monitoring practices.”

“Built a production AI-powered university marking system that automates question generation and grading from PDF course materials using a RAG pipeline (S3 + Pinecone) orchestrated with LangChain/LangGraph and deployed on AWS ECS via Docker/ECR and GitHub Actions CI/CD. Addressed a key real-world LLM challenge—grading consistency—by implementing rubric-based scoring, retrieval re-ranking, and standardized context summarization, validated against human instructors.”