What I Build

Production-grade quantitative systems with operator-level clarity.

My background spans chemical engineering, applied mathematics, data science, optimization, and ML systems. Across each role, the constant has been the same: model the system carefully, respect the constraints, and make the output useful in the real world.

I work best where research quality and operational accountability have to coexist. That usually means stochastic models, forecast engines, calibration loops, reconciliation frameworks, and reporting layers that can be trusted by both technical operators and decision makers. More recently, that same operating style has extended naturally into agentic AI and intelligent systems: LLM-backed workflows, tool-using agents, and orchestration layers that help models act safely and usefully inside production constraints.

Pricing and valuation

Pricing engines for complex financial products, calibration workflows, elasticity estimation, and margin-aware decision support.

Forecasting and uncertainty

Forecast systems with reconciliation, confidence bands, lifecycle logic, and scenario-aware outputs that survive operational use.

Risk analytics

Replay systems, payout attribution, abuse or behavior research, and analytical frameworks for explaining performance under uncertainty.

Production pipelines

Libraries and pipelines that connect research code, transactional data, monitoring, and executive reporting without losing rigor.

Agentic AI and intelligent systems

Tool-aware LLM workflows, evaluation and tracing with LangSmith, graph-based orchestration with LangGraph, and MCP-style integrations that let agents reason, call tools, and operate within real business boundaries.

Selected Private Work

High-complexity work, summarized at the right altitude.

Some of the most ambitious work in this repository is intentionally kept at a high level. The common thread across those projects is large-scale simulation, calibration, forecasting, reconciliation, and production decision support. That same private work now also includes intelligent-system patterns for orchestrating tools, structuring agent loops, tracing behavior, and making LLM outputs more auditable in production-like settings.

Monte Carlo pricing Nonlinear optimization Cohort and payout attribution Forecast confidence bands Vectorized NumPy systems SQL-heavy analytics Modular research libraries Executive-ready reporting LangChain workflows LangSmith tracing and evaluation LangGraph orchestration MCP-enabled agents

Industry Footprint

Applied math across sectors with very different operating constraints.

Since 2011, I have solved applied math and analytics problems for organizations operating in energy, finance, consulting, logistics, public sector, retail, and deep technical domains.

Chemicals
Information Technologies
Mobility

Current role: Data Scientist at FundingPips.

Earlier Public Projects

Selected public-facing work and research trail.

Data Enigma

Independent applied data science work and technical writing on analytics, AI, and visualization.

Trading Enigma

Research and system development around algorithmic trading, statistical arbitrage, and portfolio analysis.

Background

Engineering roots, quantitative range, and long-cycle technical depth.

The early engineering years shaped how I still approach quantitative work today: understand the process, formalize the constraints, build the model, and make the result operationally defensible.

Career Arc

Started as a process engineer, then transitioned into full-time data science, quantitative modeling, and production analytics.

Full CV