Senior AI/ML Architect and Engineer
Resume:
MOH IQBAL
Lead AI/ML Architect Senior ML ENGINEER ML ENGINEER DATA
SCIENTIST
********@*****.***
San Franscisco, CA 94115
PROFESSIONAL SUMMARY
Experienced AI/ML Architect with expertise in developing and deploying advanced deep learning, reinforcement learning, and multimodal AI systems. Skilled in cloud platforms (AWS, GCP, Azure), big data tools (Spark, Kafka), and MLOps frameworks (Kubeflow, MLflow). Proficient in generative AI, model optimization, and real-time analytics. Strong background in AI platform architecture and infrastructure automation, with a focus on scalability and performance. Experienced in deploying AI solutions across industries, including healthcare, finance, and e-commerce, with a commitment to responsible AI practices, such as fairness, explainability, and privacy-preserving machine learning. SKILLS
• Machine Learning & Deep Learning
PyTorch, TensorFlow, JAX, Scikit-learn, HuggingFace, CNNs, RNNs, LSTMs, Transformers, YOLO, Faster R-CNN, Multimodal Models, GANs, Diffusion Models (Stable Diffusion, DALL·E, MidJourney), VAEs, Text-to-Image Synthesis, BERT, GPT, T5, LLaMA, FLAN, LoRA/QLoRA/PEFT Fine-Tuning, Prompt Engineering, RAG Pipelines, Text Generation, Image Synthesis, Summarization, Question Answering, NER, Reinforcement Learning (RL), Multi-agent Systems, Safe Exploration, Model Parallelism (Horovod, DeepSpeed), Model Compression, Quantization, Pruning, Mixed Precision (FP16/BF16), Hyperparameter Tuning (Optuna/Ray Tune), Cross-Validation, Regularization, Anomaly Detection, Time-Series Forecasting, Predictive Modeling
• AI Systems, Data & Retrieval
Semantic/Vector Search (FAISS, Milvus, Pinecone, Weaviate), ANN, Embedding Optimization, Real-Time Inference, Streaming Pipelines
(Kafka/Kinesis), OCR & Document AI, Recommender Systems, Knowledge Graphs, Personalization Algorithms, Fraud Detection, Predictive Analytics, Deep Feature Synthesis, Automated Machine Learning (AutoML), Data Labeling & Annotation
• Agentic AI & Autonomous Systems
Reinforcement Learning (RL), Multi-agent Reinforcement Learning
(MARL), Autonomous Agents, Agent-based Simulation, Decision- Making Systems, Safe Exploration, Deep RL, Bandit Algorithms, Autonomous Robotic Systems, Real-time Agent Interaction, Learning in Dynamic Environments, Safe and Ethical AI for Autonomous Systems
• Cloud & Big Data Platforms
AWS SageMaker, GCP Vertex AI, Azure ML, Databricks, Snowflake, Apache Spark, Kafka, Kinesis, Hadoop, Redshift, BigQuery, NoSQL Databases (Cassandra, MongoDB), Data Lakes, ETL Pipelines, Scalable Data Ingestion, Cloud Storage Solutions, Streaming Data Platforms
• Responsible AI & Security
Explainability (SHAP/LIME), Bias/Fairness, Model Safety, Secure Endpoints, Ethical AI Development, Model Transparency, Trustworthy AI, Privacy-Preserving Machine Learning (Differential Privacy, Homomorphic Encryption), Fairness & Accountability in AI Models, Adversarial Robustness, AI Governance
• Software Engineering
Python, R, Java, C++, Scala, Go, Julia, SQL, Bash/Shell scripting, MATLAB, JavaScript (Node.js, React), TypeScript, Swift, Kotlin, Ruby, HTML/CSS, CUDA (GPU programming), TensorFlow.js, Rust, Spark SQL, Spark MLlib, Pandas, NumPy, SciPy, Dask, PySpark, Jupyter Notebooks, Apache Kafka, GraphQL
• Generative AI
GANs, Diffusion Models (Stable Diffusion, DALL·E, MidJourney), Text- to-Image/Video Generation, Synthetic Data Generation, DeepDream, Image-to-Text Models, AI-driven Content Creation, Natural Language Generation (NLG), Fine-Tuning LLMs for Creative and Domain- Specific Tasks, Multi-modal AI Applications
• AI Platform Engineering & Solution Architecture
Cloud-based AI Solutions (AWS SageMaker, GCP Vertex AI, Azure ML), Multi-cloud and Hybrid AI Architectures, Scalable AI Platform Design, AI Infrastructure Automation, Serverless AI Architectures, Model-as-a-Service (MaaS), End-to-End ML Pipelines, Microservices for AI, API Development, Cost Optimization for AI Workloads, Enterprise AI Integration, Kubernetes (EKS/GKE/AKS), Docker, Model Deployment (ONNX/TensorRT, Triton, KServe, FastAPI), MLOps Pipelines (Kubeflow, MLflow), CI/CD for ML
• AI Trainer & Model Deployment
Fine-tuning Domain-Specific Models, AI Model Training Pipelines, Supervised/Unsupervised Learning, Transfer Learning, Federated Learning, Model Testing & Validation, Cross-Validation, Hyperparameter Tuning, Model Interpretability, Model Versioning & Registry, Data Versioning (DVC), GPU/TPU Scaling, Model Drift Detection, Active Learning, Explainability (SHAP/LIME), Bias Mitigation, Ethical AI, Regulatory Compliance (HIPAA/GDPR), Model Safety & Security
• MLOps, Deployment & Infrastructure
MLflow, Kubeflow, Feature Stores, Model Registry, CI/CD for ML, Kubernetes (EKS/GKE/AKS), Docker, FastAPI/gRPC, KServe, Triton, ONNX/TensorRT/TVM, GPU/TPU Scaling, Model Monitoring
(Prometheus/Grafana), Airflow, Model Governance, AutoML, Real-Time Analytics & Streaming Pipelines, DevSecOps for ML Systems
• Healthcare AI & Medical ML
MONAI, DICOM, HL7/FHIR, EHR, NVIDIA Clara Imaging Workflows, Clinical NLP, HIPAA-Compliant ML Systems, Predictive Healthcare Models, Medical Imaging (Radiology, Pathology), AI for Diagnostics, Personalized Medicine, Clinical Decision Support Systems (CDSS), Medical Data Privacy, PHI-Safe Preprocessing
• Finance & E-Commerce AI
Fraud Detection, Risk Scoring, AML/KYC Models, Forecasting, Recommenders, Ranking/Search, Demand Forecasting
PROFESSIONAL EXPERIENCE
Anysphere
Lead AI/ML Architect
• Architected Cursor’s multimodal AI coding intelligence platform—LLMs, real-time code embeddings, and hybrid semantic search—boosting code- generation accuracy by 35%.
• Led development of BugBot, a multimodal debugging agent using OCR, Whisper ASR, and transformer models to analyze code, logs, and voice input.
• Designed large-scale distributed training infrastructure (PyTorch, Lightning, Ray, TPU/GPU clusters), reducing training time by 40–60%.
• Built enterprise-grade MLOps systems (MLflow, Kubeflow, CI/CD, registries), enabling weekly releases and improving reliability by 50%+.
• Engineered low-latency streaming inference (Kafka/Kinesis + FastAPI/gRPC) serving millions of daily requests.
• Developed advanced LLM fine-tuning pipelines (LoRA, QLoRA, PEFT) for code, docs, and multimodal reasoning—cutting hallucinations by 28%.
• Owned Responsible AI & security workflows: explainability dashboards, bias metrics, secure endpoints, GDPR-compliant data pipelines.
• Collaborated with infra/product teams to design microservices architectures, optimize HPC workloads, and deploy high-performance ONNX/TensorRT models.
Heavy.AI
Senior ML Engineer
• Built geospatial/time-series ML models for GPU-accelerated analytics, improving predictive accuracy by 30%.
• Designed advanced CV pipelines for satellite imagery (YOLO, segmentation, multimodal transformers) to enable high-precision terrain analysis.
• Architected high-throughput streaming pipelines (Kafka, Kinesis, Spark, TFRecord, Petastorm) for massive real-time feature extraction.
• Built LLM-powered analytics workflows using Transformers, BERT, and RAG to enable natural-language querying across dashboards.
• Deployed and optimized ML systems using MLflow, Kubeflow, ONNX/TensorRT, and Kubernetes—reducing inference latency by 45%.
• Led automation of MLOps systems: feature stores, registries, CI/CD, fairness/interpretability (SHAP), and secure enterprise APIs.
• Integrated ML workloads directly into the GPU SQL engine, enabling near real-time in-database inference and visualization. Clarifai
ML Engineer Team Lead
• Led development of production-ready CV and NLP models (CNNs, BERT, multimodal architectures) for enterprise clients, driving improvements in model accuracy and business outcomes.
• Designed scalable training pipelines using PyTorch, TensorFlow, HuggingFace, and Lightning, optimizing distributed GPU clusters and reducing training time by 40%.
• Built semantic/vector search systems with FAISS, Pinecone, and Milvus for large-scale image, video, and text retrieval, enhancing the search performance and scalability of enterprise applications.
• Implemented reproducible MLOps workflows utilizing MLflow, Kubeflow, CI/CD pipelines, feature stores, and model governance practices to ensure robust and reliable model deployment.
• Built Responsible AI workflows, integrating explainability (SHAP/LIME), bias mitigation, and secure endpoints, ensuring compliance with GDPR and HIPAA regulations.
• Mentored engineers and collaborated cross-functionally to deploy microservices, optimize ONNX/TensorRT inference, and deliver scalable ML applications in production environments.
Eyeris Technologies
Data Scientist
• Developed predictive models for customer behavior analysis, improving retention by 20% through feature engineering and statistical modeling techniques.
• Performed data cleaning, transformation, and exploratory data analysis (EDA) on large datasets, identifying key insights that influenced strategic business decisions.
• Collaborated with cross-functional teams to design and deploy machine learning algorithms for real-time recommendation systems, resulting in a 15% increase in user engagement.
• Built and optimized machine learning pipelines, improving model training efficiency by 30% and reducing deployment time by automating data preprocessing and model evaluation.
PROJECTS
AI Coding Intelligence Platform – Anysphere
• Architected end-to-end AI pipelines for Cursor with multimodal LLMs and hybrid semantic search, boosting code-generation accuracy by 35%.
• Created BugBot using OCR and Whisper ASR for automated debugging, reducing developer troubleshooting time by 40%.
• Built distributed training infrastructure with real-time streaming inference, cutting deployment time by 50%. Geospatial & Enterprise Analytics – Heavy.AI
• Developed predictive ML models for geospatial and time-series data, improving forecasting accuracy by 30%.
• Designed CV pipelines for satellite imagery with segmentation and transformers for object detection. 10/2022 – Present
08/2019 – 09/2022
06/2017 – 07/2019
02/2014 – 05/2017
• Implemented LLM-powered analytics workflows, reducing inference latency by 45%. Enterprise AI & Vector Search – Clarifai
• Built scalable semantic/vector search systems (FAISS, Pinecone, Milvus) for enterprise AI applications.
• Developed NLP models for knowledge extraction and contextual search across large datasets. In-Cabin AI Systems – Eyeris Technologies
• Developed CV models for in-cabin driver monitoring and safety detection.
• Deployed models with ONNX/TensorRT for efficient edge inference in automotive systems. CERTIFICATES
Google Cloud Professional Machine Learning Engineer PyTorch Advanced Certificate
Databricks Professional ML & Data Engineer
HuggingFace Transformers Certification
AI Ethics and Governance
EDUCATION
Computer Science
University of houston
Contact this candidate