Realtime feed · AI/ML Engineer

|

I design and ship end‑to‑end AI systems — agents, data/ML pipelines, and operational interfaces—built for reliability, observability, and speed.

Years shipping: 5+ in AI/ML & platforms
Founder builds: 2 ventures launched
Current mission: AI/ML Engineer @ ValueMomentum

Agent Nodes

47 active

Edge signals

Live/142

Trust vector

0.997

Telemetry Grid

Streaming highlight packets straight from the build loop.

Multi-agent orchestration

50% faster

Multi-agent workflow shaved insurance handling time in half.

On-prem inference

142 tokens/s

On-prem LLM cluster keeps PHI private and eliminates API cost spikes.

Async pipelines

2× throughput

Python asyncio queues doubled AI service throughput.

Mission Archive

Selected deployments blending AI autonomy with rich interfaces.

2024 – PresentLangGraph/LangchainCrewAIAzureAWSOllamavLLMViTLayoutLMv3CNNs

ValueMomentum · AI/ML Engineer

Built LangGraph-based multi-agent orchestration framework, on-prem deployments, LayoutLMv3 document intelligence, and ViT pipelines for geospatial risk. Experimented with various LLMs, ML architectures and AI frameworks. Led AI efforts and enterprise platform development.

2024SparkCI/CDPython

Swiss Re · Data Engineer Intern

Optimized Spark jobs across 20 TB, hardened CI/CD with unit tests, and shipped a structured logging library that halved debugging effort.

2023 – 2024Reinforcement LearningPyTorchVR

C2SMART @ NYU · Research

Prototyped Deep Q-Learning agents and quaternion time-series forecasting for immersive VR simulations.

2019 – 2022GCPFlaskKubernetes

Immigreat · Founding Engineer

Architected serverless backends, Bitbucket CI/CD, and Kubernetes Autopilot migrations powering immigration workflow products.

Systems Matrix

Capabilities mapped to velocity + confidence vectors.

LLM + Generative AI Systems

Langchain, LangGraph, CrewAI, and guardrailed LLM workflows and multi-agent systems tuned for domain precision.

Data & ML Engineering

PyTorch, Spark, and async Python services delivering high-scale inference.

Product-Focused Leadership

Translating research breakthroughs into measurable business outcomes.

Project Constellations

Curated repos grouped by craft area—AI tooling, open-source libraries, full-stack systems, generative apps, academic research, and platform ops.

EzLangStudio · FastAPI + LangGraph + Next.js

EzLangStudio

Full-stack studio that lets teams sketch multi-agent LangGraph flows, run and test workflows, and export to production-ready code.

Developer cockpit for visually assembling LangGraph agents with a FastAPI brain and Next.js interface.

Visual workflow builder with version history, duplication, and validation tools.
FastAPI service that persists LangGraph DSL, streams run updates over SSE, and coordinates exports.
LangGraph runtime with pluggable checkpointing (SQLite, Postgres, in-memory) and optional Redis integrations.

LangGraphFastAPINext.js

Experimental Lab

Concept modules under active iteration.

LangGraph Cortex

Agents parsing claims queues with auto-escalation hooks.

CrewAI Fabric

Autodocs + PR review rituals baked into CI pipelines.

Multi-agent Orchestrator

Custom production-grade agentic framework wrapping LangChain/LangGraph.

Signal Stream

Live packets from the lab cluster, decoded for humans.

LANGGRAPH

LangGraph agents reconciled claim bundles + cut manual triage 50%.