← sneha hassan
3 roles · research, product engineering, ML

Experience

From shipping to riders at Ather Energy to LLM research at Carnegie Mellon.

FEB 2026 - PRESENT · PITTSBURGH

Carnegie Mellon University

Research Engineer (GenAI)
  • Fine-tuned a majority-voting ensemble of three domain-specialized LLMs on curated bio-imaging datasets; implemented consensus-based iterative training evaluated across 296 bioinformatics benchmarks, achieving a 25% accuracy improvement.
  • Architected and deployed a custom 8B-param LLM with tool-calling for Cryo-Electron Tomography using PyTorch and transfer learning, automating cellular annotation that took researchers 8 hours of manual labeling per dataset.
  • Built an async evaluation infrastructure with automated grading across bioinformatics benchmarks; optimized inference with batch processing and caching, cutting per-sample time from 45s to 12s (73% faster).
MAY 2025 - AUG 2025 · MT. LAUREL, NJ

Telvue Corporation

Software Developer Intern
  • Created an event-driven push notification system using Rails microservices and Firebase Admin SDK with pub/sub triggers and retry logic, increasing mobile user engagement by 40% across 13K+ users.
  • Led a 4-person cross-functional team (engineering, design, mobile) to unify notification UX across iOS and Android, reducing design to production integration time from 3 weeks to under 2 weeks.
FEB 2023 - NOV 2024 · BANGALORE

Ather Energy

Software Developer
  • Architected a B2B fleet monitoring dashboard with real-time IoT telemetry (remote immobilizer, GPS, diagnostics) using React and Firebase, adopted by 3 enterprise clients managing 200+ vehicles.
  • Engineered and shipped a full-stack PWA using Spring Boot (Java) and React Native for carbon-footprint tracking, serving 50K+ DAU and reducing end-to-end tracking latency by 25% through API response caching.
  • Trained and deployed an ML pothole detection model (92% accuracy) using Python signal processing on accelerometer data, driving a 35% increase in safety-feature adoption across the rider base.
  • Owned a cross-team test automation platform using Java and Flutter, increasing automated test coverage from 32% to 78% and cutting QA cycle time by 45% across 4 engineering squads.
  • Streamlined 12 legacy Firebase Cloud Functions (Node.js) and built custom SQL automation scripts, reducing average API response time by 20% and data processing overhead by 10%.
  • Developed 6 RESTful APIs in Spring Boot powering a favorite-locations feature for the Ather app (100K+ users), cutting location-search latency by 40% with indexed queries and Redis caching.
  • Built a full-stack campaign PWA using React frontend and Java backend, enabling real-time user participation across 10+ cities and driving 25K+ campaign sign-ups in the first month.
Stack of choice
PythonTypeScriptReact / Next.jsReact NativeJava / Spring BootNode.jsFlaskPyTorchTensorFlowLLM fine-tuningLangChain / RAGOpenCVFirebaseAWS / GCPDockerPostgreSQLMongoDBRedis