Job Description
The context: Shadowfax just closed its most profitable quarter ever. โน1,358 Cr revenue (+65% YoY), a fifth straight quarter of 60%+ growth, and all-time-high PAT of โน65 Cr. And AI here is in production, not in slideware: our delivery-partner copilot handles ~16,000 conversations a day with ~97% resolved without human intervention, and Vision AI catches ~40% of mismatched reverse pickups at the doorstep at ~35x lower inference cost than a frontier model.
The CFO's office runs the same way. Agentic reconciliation, LLM-assisted anomaly detection, and self-refreshing dashboards already run our revenue-assurance workflows. We are hiring the engineer who takes this system 10x further.
๐๐ฏ๐ผ๐๐ ๐๐ต๐ฒ ๐ฟ๐ผ๐น๐ฒ
You will be the AI engineer inside Business Finance and Revenue Assurance: one engineer, working with AI, producing the output of a team, at public-company accuracy standards. The systems you build protect revenue across 1 Cr+ shipments a month.
๐ช๐ต๐ฎ๐ ๐๐ผ๐'๐น๐น ๐ฏ๐๐ถ๐น๐ฑ
โ Agentic AI workflows (Claude, GPT, or equivalent) that reconcile 1 Cr+ shipments monthly
โ LLM-assisted anomaly detection that flags non-compliance before month close, not after
โ ML models for revenue-leakage and fraud detection: time-series anomaly detection, transaction matching, variance decomposition
โ The finance data layer: SQL/Python ETL pipelines over OMS, TMS, and billing-system extracts, validated and reconciled to source
โ Evaluation harnesses for every AI workflow (golden cases, regression checks) so no unverified number reaches leadership
โ End-to-end reconciliation automation: transaction matching, variance detection, automated settlement workflows
๐ฌ๐ผ๐ ๐๐ถ๐น๐น ๐ฏ๐ฒ ๐บ๐ฒ๐ฎ๐๐๐ฟ๐ฒ๐ฑ ๐ผ๐ป
โข Rupees recovered and leakage prevented by systems you build
โข Hours of manual finance work eliminated
โข Accuracy of AI outputs in production: eval pass rates, error budgets
โข Speed from leadership question to verified answer
๐ช๐ต๐ผ ๐๐ต๐ผ๐๐น๐ฑ ๐ฎ๐ฝ๐ฝ๐น๐
โข 2-4 years building production ML or AI systems; you have shipped something AI-powered that people actually use
โข Proficiency in machine learning and pattern recognition, including designing, training, and evaluating models for complex business problems; NLP for document understanding and query-based analytics
โข Expert Python plus deployment (FastAPI, Docker); advanced SQL and feature engineering
โข Hands-on LLM work: Claude / GPT / Gemini APIs, prompt engineering, RAG, agentic workflows or MCP
โข B.Tech / M.Tech in CS, Data Science, or AI-ML
โข Logistics, fintech, or high-volume transactional domain experience preferred
โข Accuracy obsession: an unverified number is a defect, not a draft
๐ช๐ต๐ฎ๐ ๐๐ต๐ถ๐ ๐ฟ๐ผ๐น๐ฒ ๐ถ๐ ๐ป๐ผ๐
โข Not AI research: this is applied AI on live financial data, judged by rupees recovered and hours saved
โข Not a support seat: you own systems end to end, from pipeline to production to evaluation
โข Not a prompt-only role: you ship code that runs unattended
๐ช๐ต๐ ๐ท๐ผ๐ถ๐ป
โข Build agentic AI in mission-critical finance at a listed company, with real P&L data from day one
โข Direct CXO exposure: your systems feed pricing, margin, and commercial decisions at the leadership table
โข A team already operating AI-first, where the path from Associate to Director has been walked in two years
๐๐ผ๐บ๐ฝ๐ฒ๐ป๐๐ฎ๐๐ถ๐ผ๐ป: Competitive CTC, benchmarked to top quartile. Fixed plus performance variable.
#Hiring #AIEngineer #MachineLearning #BusinessFinance #Bangalore #Shadowfax