About Turing
Based in San Francisco, California, Turing is the world’s leading research accelerator for frontier AI labs and a trusted partner for global enterprises deploying advanced AI systems.
Turing supports customers by accelerating frontier research with high-quality data and expert contributors, and by helping enterprises deploy AI systems that deliver measurable business impact.
About the Role
We are seeking an engineer or computer scientist who can bridge Vedic Astrology (Jyotish) and computational astronomy.
Your primary responsibility is to build Python-based verification pipelines that compute and validate planetary positions and related Jyotish elements using astronomical libraries such as pyswisseph (Swiss Ephemeris) and/or Skyfield. Your scripts will serve as ground-truth generators and validators for an AI evaluation dataset.
You should be equally comfortable debugging an ephemeris call as you are reasoning about Sayana vs Nirayana coordinate systems and the role of Ayanamsa transformations.
Key Responsibilities
- Write Python verification scripts using pyswisseph and/or Skyfield to compute and verify:
- Planetary longitudes (geocentric, apparent) for date/time/location
- Nirayana (sidereal) longitude conversion via specified Ayanamsa (Lahiri, Krishnamurti, Raman, etc.)
- Ascendant (Lagna) degree computation for birth data
- House cusp calculations (Equal, Placidus, Koch, Whole Sign)
- Nakshatra and Pada determination from computed longitude
- Sunrise/Sunset times for given coordinates and dates
- Build automated test harnesses to validate batches of Kundali computation problems against ephemeris-derived ground truth.
- Create 10–15 code-verified evaluation problems per month, with algorithmically generated answers.
- Document discrepancies across ephemeris sources and Ayanamsa systems with quantified error margins.
- Develop utility scripts for Panchang calculations:
- Tithi (Sun–Moon elongation)
- Nakshatra (Moon longitude)
- Yoga (Sun + Moon longitude)
- Karana (Tithi subdivisions)
- Build a verification pipeline usable by non-coding reviewers for spot-checking manually created problems.
- Collaborate with classical Jyotish contributors to translate traditional methods into code and identify edge cases.
Required Qualifications
- Bachelor’s or Master’s degree in Computer Science, Software Engineering, Physics, Astronomy, or Mathematics.
- Minimum 3 years of professional Python development experience.
- Proficiency with at least one astronomical computation library:
- pyswisseph (Swiss Ephemeris bindings): e.g., `swe_calc_ut()`, `swe_houses()`, `swe_set_sid_mode()`
- Skyfield: loading ephemerides (e.g., `de421.bsp`), coordinate transforms, timescale handling
- Understanding of Sayana (tropical) vs Nirayana (sidereal) systems and how Ayanamsa bridges them.
- Familiarity with astronomical time systems (UT, UTC, IST, Julian Date) and coordinate systems (ecliptic/equatorial).
- Basic Jyotish concepts (Rashi, Nakshatra, Lagna, Bhava, Graha) sufficient to interpret computations.
- Strong documentation skills: write reproducible, well-commented code explaining both astronomy and Jyotish relevance.
Preferred Qualifications
- Contributions to open-source Jyotish or astronomy projects.
- Experience building Jyotish applications (web/mobile/desktop horoscope tools).
- Knowledge of JPL ephemerides (DE430/DE431) and precision characteristics.
- Understanding of Surya Siddhanta algorithms vs modern ephemeris.
- Published writing on computational astrology.
- Familiarity with year-by-year differences among Lahiri/Krishnamurti/Raman Ayanamsa values.
- Experience with geodetic calculations (lat/long → local sidereal time).
Tools & Technologies
- Python 3.8+ (NumPy, pandas, datetime, pytz)
- `pyswisseph` (Swiss Ephemeris bindings)
- `skyfield` (JPL ephemeris positional astronomy)
- `astropy` (optional: coordinate transforms and time handling)
- Jupyter Notebook for verification and documentation
- Git/GitHub for version control
- Jyotish software (e.g., Jagannatha Hora) for cross-validation
Engagement Details
- Remote contract engagement.
- Deliverable-based: 10–15 code-verified evaluation problems per month.