About the Role
Job Summary
We are looking for a versatile Full Stack Python Developer to take ownership of a production-grade crypto market-structure analysis platform. The system combines real-time OHLCV data ingestion, algorithmic swing/trend detection (BoS, CHoCH, Wyckoff), AI-generated market summaries, and an interactive Streamlit front end — all backed by a Django REST API and a multi-agent LLM pipeline.
You will be the primary engineer: maintaining and extending existing detection algorithms, improving the AI summarization pipeline, hardening authentication and integrations, and shipping new visualizMaintain and improve the Django 5 async REST API (django-ninja), domain-separated apps (AI, chat, stock, auth, chart, users) and Pydantic schema layer.
Responsibilities
Extend swing-point and trend detection algorithms (BoS/CHoCH, Wyckoff zones) built on pandas DataFrames — improving accuracy, coverage, and performance.
Develop and tune the LLM/AI pipeline: LangChain agents, LangGraph orchestration, OpenAI integration, and Langfuse prompt tracing/observability.
Build and refine the Streamlit interactive UI: Plotly candlestick charts, overlays, session state management, and browser-side token persistence.
Own the Celery + Redis background task layer, async job scheduling, and health-check infrastructure.
Manage exchange API integrations (Bybit public API patterns), token-refresh auth flows, and REST ingestion helpers.
Maintain Docker-based dev and production infrastructure (multi-stage builds, compose stacks, MinIO/S3, uWSGI).
Write and expand the pytest test suite; enforce code quality with ruff, pre-commit hooks, and type annotations.
Qualifications
Python 3.11+ — strong command of modern Python: type hints, dataclasses, enums, packaging (pyproject.toml / Poetry / UV).
Django (v5) — async patterns, modular app design, django-ninja or DRF, ORM with custom querysets/managers.
Pandas & NumPy — time-series manipulation, rolling/EWM operations, index-based masking on OHLCV data.
Streamlit — production deployments, session state, component lifecycle, local storage integration.
LLM / AI engineering — LangChain, agent/tool design, structured outputs, prompt engineering; OpenAI API experience required.
Plotly — candlestick charts, graph_objects overlays, interactive financial visualizations.
REST API design — token-based auth, OAuth/refresh patterns, Pydantic request/response schemas.
Celery + Redis — background task design, scheduling (celery-beat), broker/result-backend configuration.
Docker — multi-stage Dockerfiles, compose stacks, service orchestration.
Testing & quality — pytest, factory_boy, ruff linting, pre-commit, CI pipelines.
Trading/market-structure basics — knowledge of or willingness to learn swing highs/lows, Break of Structure (BoS), Change of Character (CHoCH), and Wyckoff methodology.
NICE TO HAVE
- LangGraph and LangFuse (or equivalent LLM observability stack) hands-on experience.
- Bybit or other crypto exchange API integration experience.
- Pydantic v2, OpenTelemetry, and Sentry instrumentation in a production setting.
- MinIO / S3-compatible object storage configuration.
- PostgreSQL with psycopg3, async queries, and geospatial extensions (GDAL/GEOS).
- RS256 JWT key management and cryptography-library-level security work.
- Experience serializing dataframes as LLM context (CSV-based prompt construction).
TECH SNAPSHOT
Backend: Python 3.13, Django 5, django-ninja, Celery, Redis, PostgreSQL, uWSGI
Frontend: Streamlit, Plotly (graph_objects), streamlit-local-storage
AI / LLM: LangChain, LangGraph, OpenAI, Langfuse, prompt engineering
Data: Pandas, NumPy, ta (technical analysis), Pydantic
Infra: Docker, MinIO/S3, OpenTelemetry, Sentry, JWT (RS256)
Quality: pytest, factory_boy, ruff, pre-commit, Poetry / UV
Pay: Up to $183,000.00 per year
Work Location: Remote