Apoorv Awasthi

With 11 years as an Artificial Intelligence Engineer, I approach challenges with determination and a strong logical framework, always seeking to achieve mastery in my craft. My focus is on delivering pragmatic, accountable solutions, leveraging resourcefulness to build impactful AI systems.

Experience

Artificial Intelligence Engineer

Outlier AI · May 2026 – Present

Not yet confirmed

• • Authored 300+ evaluation tasks assessing AI-generated TypeScript and Python against production-quality standards on large open-source codebases, designing the grading rubrics and criteria used to score code-generation models.
• • Produced PR-level cross-file context and reference solutions used in model fine-tuning and RLHF alignment pipelines, improving code-generation accuracy on real-world repositories.

O

Software Engineer - AI Evaluation (Contract)

Outlier AI, Remote · May 2026 – Present

Not yet confirmed

• Authored 300+ deep technical tasks designed to challenge LLM coding models on complex repository-level changes, including multi-file state management and API flows requiring cross-file data and logic.

Software Developer

Spring Money · May 2025 – May 2026

Not yet confirmed

• • Designed and implemented a metadata-driven configuration platform and REST APIs powering 200+ configurable workflows across 5 client instances, enabling client-specific changes without releases and reducing client-driven redeployments by 80%.
• • Built a configurable business-rule engine supporting 100+ client-defined rules across 18 operators and 6 evaluation categories, driving validation and workflow decisions across 120+ workflows.
• • Built backend services in Python/Frappe and a Next.js/TypeScript BFF for a financial planning platform serving web and mobile clients, reducing portfolio and account dashboard API latency from 14s to 3s.
• • Designed a normalized financial data model spanning customers, accounts, goals, assets, liabilities, and cash flows across 20+ asset classes, enabling cross-entity aggregation and portfolio analytics.
• • Owned 4 core fintech workflows end-to-end — onboarding, financial data gathering, financial planning, and goal mapping — from API design and data modeling through production, onboarding 200+ client accounts across 11 advisors.
• • Improved platform efficiency through two-tier caching, reducing redundant requests by 25%, and automated CI/CD workflows with GitHub Actions and AWS EC2, cutting deployment process time by nearly 30%.
• Technologies Used: Python, Frappe, TypeScript, Next.js, React, React Native, REST APIs, MySQL, React Query, GitHub Actions, AWS EC2, Turborepo, Jest, Vitest, Zod

S

Software Development Engineer

Spring Money, Pune, India · May 2025 – Apr 2026

Not yet confirmed

• Engineered a metadata-driven configuration service that generates 200+ client-customizable workflows from persisted schemas, cutting redeployments by 80% and new client onboarding to 15 days.
• Built a configuration-driven rules engine with 18 operators and 6 action types powering form behavior across 120+ client workflows, and a server-side state machine in Python resolving next steps from persisted progress, replacing hardcoded branching with declarative rules.
• Reduced backend API latency by 77% in Python by decomposing 10-12 sequential computations into 5-6 parallel endpoints and caching shared computed values in Redis.
• Implemented two-tier caching with Redis-backed server reads and React Query request deduplication, cutting redundant API calls by 25% and easing backend load.
• Owned 4 core fintech workflows end-to-end (onboarding, data gathering, financial planning, goal mapping), from normalized data modeling across 20+ asset classes to REST API design and production, serving 200+ portfolios.
• Automated GitHub Actions pipelines for branch syncing, unit tests, and AWS EC2 redeployment, cutting deployment time by 30% after consolidating the frontend apps into a Turborepo monorepo.

F

Software Engineer

Fibiso Technologies · Jun 2024 – Apr 2025

Not yet confirmed

• • Optimized initial bundle size via route-based code splitting and lazy loading, cutting time-to-interactive to under 3s on large data grids.
• • Developed reusable React components and a normalized Redux state layer handling 500+ live aircraft schedules with memoized selectors, minimizing re-renders.

F

Software Development Engineer | Client: Ramco Systems

Fibiso Technologies, Bengaluru, India · Jun 2024 – Apr 2025

Not yet confirmed

• Virtualized the heaviest grids with windowed rendering, server-side filtering, and route-based code splitting, cutting time-to-interactive by 50% (5s to 2.4s).
• Added client-side caching in TypeScript that diffs edited rows against cached state and invalidates only affected entities, replacing full-collection refetches after every edit, cutting repeat API calls by 25%.
• Built a Node.js middleware layer transforming and normalizing data from multiple backend APIs into frontend-specific models for React scheduler views, handling 5000+ crew, flight, and booking records with server-side sorting.
• Rebuilt 10+ screens of Ramco Systems' aircraft management platform on a shared data layer, unifying aircraft schedules, crew rosters, flight availability, and bookings.

F

Software Developer

Fibiso Technologies · Jun 2024 – Apr 2025

Not yet confirmed

• • Designed and optimized a React frontend architecture for a real-time aviation data system handling 500+ live aircraft schedules.
• • Developed reusable React components and a normalized Redux state layer with memoized selectors, minimizing unnecessary re-renders across large data grids.
• • Reduced initial JavaScript bundle size through route-based code splitting and lazy loading, bringing time-to-interactive to under 3 seconds on large data grids.
• • Built performance-focused UI components for high-frequency, real-time data, optimizing rendering and state updates for continuously changing aircraft schedules.
• Technologies Used: React, JavaScript, Redux, HTML, CSS

G

Software Developer Intern

Glocol Networks, USA · Aug 2023 – Dec 2023

Not yet confirmed

• Built analytics dashboards for a California rail operator, aggregating inter-station passenger flow across routes carrying 10,000+ daily commuters, used to assess capacity and schedule added service.
• Wrote Bluetooth collection software for Raspberry Pi 0 W units aboard trains, tuning scan intervals and duty cycles to balance device capture against power draw and duplicates, feeding detections to an AWS cloud pipeline.

G

Software Developer

GLOCOL INC · Aug 2023 – Dec 2023

Not yet confirmed

• • Rebuilt Vue.js dashboards for real-time passenger information and occupancy monitoring, deployed across 4+ cities and serving 800+ users.
• • Developed real-time IoT data pipelines and visualizations for monitoring train passenger occupancy, connecting Raspberry Pi-based scanning devices to cloud services and operational dashboards.
• • Built APIs and backend integrations to ingest, process, and expose real-time train occupancy data from IoT scanner devices.
• • Shipped real-time passenger occupancy visualizations supporting transportation operations and impacting 10,000+ commuters.
• Technologies Used: Vue.js, JavaScript, AWS Lambda, Amazon DynamoDB, AWS IoT Greengrass, API Gateway, Raspberry Pi, REST APIs

G

Software Engineer Intern

Glocol Networks · Aug 2023 – Dec 2023

Not yet confirmed

• • Rebuilt Vue.js dashboards deployed across 4+ cities, serving 800+ users.
• • Shipped real-time IoT visualizations impacting 10,000+ commuters.

B

Full-stack Developer

Buisness Web Solutions · Aug 2022 – Oct 2022

Not yet confirmed