AI Engineer

1. M T CNG VIC

- Thit k v pht trin h thng AI to nh, phc v cho cc tc v nh chuyn i phong cch, chnh sa trang phc, nhn vt, hoc to hnh nh mi t input text/image.

- Nghin cu, hun luyn v ti u ha m hnh AI (Stable Diffusion, StyleGAN, ControlNet, Deep Image Prior, v.v.) ci thin tc x l v cht lng u ra.

- Xy dng pipeline AI hon chnh, t ng ha quy trnh to nh v ti u quy trnh inference.

- Tinh chnh m hnh AI (fine-tuning, LoRA, DreamBooth, Textual Inversion, v.v.) ph hp vi tng nhu cu c th.

- Ti u ha hiu sut v trin khai AI trn mi trng cloud (AWS, Google Cloud, Azure) hoc my ch on-premise c GPU.

- Pht trin API hoc dch v AI tch hp vo h thng backend ca cng ty.

- Phi hp vi nhm k thut v sn phm a AI vo cc sn phm thc t.

- Cp nht v nghin cu cng ngh mi, m bo cng ty lun i u trong lnh vc AI to nh.

- Xy dng AI Chatbot chuyn bit cho mt s lnh vc, tch hp vo app ecosystem.

- Thit k pipeline RAG (Retrieval-Augmented Generation) t d liu, product docs.

- Fine-tune hoc orchestrate LLM (OpenAI, Llama3, Claude, vLLM) cho hi thoi hng dn, t vn.

- Trin khai LLM-based agent phc v cc workflow: phn tch hnh nh, gi routine, h tr ngi dng.

- Phi hp cht ch vi product, backend v frontend a AI vo sn phm c ngi dng tht.

- Theo di v cp nht cc cng ngh mi v multimodal foundation models, RAG, v image generation.

II. YU CU NG VIN

K nng chuyn mn:

- Ti thiu 3+ nm kinh nghim lm vic vi AI/Deep Learning, c bit l Generative AI v Computer Vision.

- Kin thc su v m hnh AI to nh, bao gm Diffusion Models, GANs, VEst, NeRF, ControlNet.

- Thnh tho Python v cc th vin AI/ML: PyTorch, TensorFlow, Diffusers, OpenCV.

- Kinh nghim hun luyn v fine-tuning m hnh AI trn dataset ln.

- Hiu bit v x l hnh nh nng cao, bao gm inpainting, out painting, texture transfer, super-resolution.

- Kinh nghim trin khai m hnh AI trn cloud (AWS, GCP, Azure) hoc ti u trn GPU (CUDA, TensorRT).

- C kh nng xy dng API AI hoc microservices tch hp vo cc h thng ln.

u tin:

- tng lm vic vi cc nn tng AI to nh nh Stable Diffusion, Mid Journey, DALLE, Runway.

- C kinh nghim vi C++/Rust cho ti u ha hiu sut inference.

- trin khai sn phm AI vo thc t, c bit l trong lnh vc game, qung co, hoc sng to ni dung.

- C bng Thc s/ Tin s v AI, Machine Learning, hoc lnh vc lin quan.

- C portfolio hoc GitHub vi cc d n AI to nh n tng.

III. QUYN LI

Lng net: 30-60M

- Ph cp n tra 500000, thu tr 200000, i li 100000;

- c lm vic v training vi Tin S AI ang lm vic ti Tesla M

- Vn ha cng ty: c sch, tp th dc hng ngy;

- Lm vic trong mi trng tr trung, thn thin v hng n vic pht trin con ngi;

- Thi gian lm vic: T2 - T6, ngh 3 th 7, 1 th 7 Creative day v 4 ch nht.

(8h30 - 18h, ngh tra 1h30p)

- Thng cc ngy ngh l, thng tt (1-5 thng lng);

- Xt duyt thng tour du lch 6 thng/ln;

- Review lng 2 ln/nm; (iu kin 6 thng chnh thc)

- Trang b thit b hin i, laptop, mn hnh ln;

- Ch BHXH, BHYT, BHTN theo lut lao ng;

- n nh / ung min ph;

- Du lch hng nm, tham gia lin hoan hng thng;

- Tham gia vo CLB Bng ca cng ty hng tun.

- Quyn li dnh cho thnh vin lm >1 nm: Khm sc khe tng qut 1 nm/ln, Cp vn mua nh, mua xe, th tp gym, yoga

IV. THNG TIN NG TUYN

Gi CV vo mail: [Confidential Information]

• a ch: To The Nine, s 9 Phm Vn ng, Mai Dch, Cu Giy, H Ni
• FB cng ty: https://www.facebook.com/xgamestudio
• Web: http://xgamestudio.com/