Senior ML Engineer

We’re developing an AI-powered SaaS CRM platform tailored for the tax services industry. Our mission is to save time, reduce manual errors, and enable smarter workflows by turning unstructured financial documents into actionable insights.

To help us get there, we’re looking for a hands-on Senior Machine Learning Engineer to lead the design and implementation of intelligent systems that bring clarity to document chaos. From research and rapid prototyping to production-grade deployment — you’ll be at the center of our machine learning efforts.

This is a rare opportunity to apply your expertise in OCR, NLP, and Large Language Models in a domain with meaningful, real-world impact: helping tax professionals work faster, more efficiently, and with greater accuracy.

🚀 What You’ll Do

• Lead the ML lifecycle – Scope problems, evaluate solutions (OCR, traditional NLP, LLMs/RAG, heuristics), deploy to production, and monitor performance.

• Own the data strategy – Define what data to collect or synthesize, and implement feedback loops for continuous model improvement.

• Productionize ML prototypes – Evolve notebooks into robust, well-tested Python services with CI/CD and clean interfaces.

• Define evaluation frameworks – Build automated benchmarks and dashboards to track model quality, latency, and cost. Communicate trade-offs clearly.

• Drive prompt engineering & fine-tuning – Craft effective prompts, design retrieval pipelines, and apply techniques like LoRA or adapters.

• Mentor and collaborate – Support teammates’ growth and align ML work with product priorities and long-term business goals.

🧠 Projects You’ll Work On

We're building a document intelligence platform that powers the next generation of SaaS tools for tax professionals. You'll contribute to:

• Analyzing previous tax returns to predict required documents for the current year.

• Classifying and organizing customer-uploaded documents using OCR and LLMs.

• Generating smart checklists and action summaries to streamline tax preparation.

• Continuously improving model accuracy with real user feedback loops.

✅ You’re a Great Fit If You Have

• 5+ years of experience building and deploying ML systems, especially in NLP or applied ML.

• Hands-on experience solving end-to-end real-world problems — from data ingestion and modeling to deployment and measurable results.

• Strong grasp of modern NLP techniques: embeddings, retrieval systems, generative models, prompt engineering, fine-tuning, and evaluation.

• Solid engineering foundations: Python, Docker, cloud GPUs, CI/CD, testing practices.

• A pragmatic, metrics-driven mindset — you understand how to balance accuracy, latency, and cost.

🌟 Nice to Have

• Experience with document-focused ML in domains like legal, finance, insurance, or healthcare.

• Background in OCR or document parsing, especially scanned or semi-structured forms.

• Familiarity with tools like docTR, Tesseract, LayoutLM, or multimodal LLMs.

💡 Why Join Us

• Make an Impact – You’ll build ML systems used by thousands of professionals in a critical domain.

• True Ownership – Influence the full ML stack — from data strategy to deployment — not just model tuning.

• Modern Tech, Real Problems – Work with cutting-edge methods in OCR, LLMs, prompt engineering, and retrieval systems.

• Autonomy & Responsibility – Minimal bureaucracy, maximum ownership. Your ideas will shape product direction.

• Tight Collaboration – Engineers, PMs, and domain experts work side-by-side with shared goals.

• Growing Market – We're operating in the high-value U.S. tax tech space with strong traction and real users.

🧪 Interview Process

1. HR Call (1 hour) – Meet with HR to align expectations and answer questions.

2. Technical Task (up to 4 hours) – Solve a real-world ML problem on your own time.

3. Technical Interview (1 hour) – Discuss your task and explore core ML knowledge.

4. System Design Interview (1 hour) – Dive deeper into how you'd architect real systems.

5. Behavioral Interview (1 hour) – Final conversation with our CTO to align on values, vision, and collaboration style.

We aim to complete the process within 1–3 weeks.

Join us in shaping the future of document intelligence in fintech — your work will directly impact thousands of professionals and their clients.