Henry Hecht

Quantitative Trader - Kalshi - Remote

(2026-02)

Grew a personal prediction-market trading account from $1K to $80K over approximately four months by designing and deploying automated trading infrastructure built around real-time event forecasting.
Designed and maintained a production software platform composed of distributed Python services, event-driven pipelines, streaming inference, and automated execution infrastructure processing real-time market data.
Built a real-time trading platform processing hundreds of live speech segments per event with automated inference and order execution.
Built a backend capable of monitoring 50+ concurrent markets while executing trades through REST/WebSocket APIs.
Developed forecasting models using fine-tuned language models, survival analysis, and probability calibration to estimate event outcomes from speaker-specific transcripts.
Engineered risk and analytics infrastructure with fractional Kelly position sizing, cross-market correlation modeling, per-strategy PnL attribution, and backtesting framework accounting for fees, slippage, and resolution edge cases.

Software Engineer - Aspen Dental - Chicago, Ill.

(2025-08 - 2026-02)

Developed internal web interfaces for configuring, monitoring, and debugging generated MCP servers.
Built TypeScript-based frontend tools enabling engineers to inspect generated APIs and integration status.
Integrated frontend applications with FastAPI/Node.js backend services through REST APIs.
Collaborated with product managers and engineering teams during sprint planning to prioritize AI platform initiatives.
Participated in peer code reviews and architectural discussions to improve maintainability and reliability.
Worked cross-functionally with platform and infrastructure engineers to deploy secure healthcare-compliant AI services.
Built an automated pipeline that parsed 1,000+ internal APIs and generated production MCP servers, reducing new integration effort by approximately 90%, saving 20–30 engineering hours per integration.
Implemented DSPy-based optimization pipeline for tool descriptions, improving tool-selection accuracy from 60% to 95%.
Implemented healthcare-compliant architecture with Model Armor, Grafana, PII redaction, and audit trails.

Contributor - Outlierai - Remote

(2024-02 - 2025-08)

Produced 500+ RLHF training pairs while maintaining 95%+ reviewer acceptance
Analyzed Python AST and control flow graphs from large open-source Python codebases including NumPy, Django, FastAPI to generate adversarial code comprehension queries testing LLM reasoning on decorators, metaclasses, and async patterns.
Developed weighted scoring rubrics with quantitative metrics for code correctness, edge case handling, and algorithmic complexity, enabling automated quality assessment of AI-generated Python solutions.

Integration Engineer - Datamotion Inc - Morristown, NJ

(2023-08 - 2024-04)

Designed and implemented a retrieval-augmented generation (RAG) platform using Ray, LangChain, Mistral, pgvector, and PostgreSQL to power context-aware search across Google Workspace emails and documents.
Built scalable ingestion pipelines for enterprise content, including document parsing, embedding generation, metadata extraction, and vector indexing.
Developed Python backend services and REST APIs supporting semantic retrieval and AI-powered document workflows for internal applications.
Optimized retrieval quality through vector search tuning, prompt engineering, and embedding evaluation while collaborating with cross-functional engineering teams to deploy production AI features.

About